Customer service system dialogue scheduling method and device, electronic equipment and storage medium

Abstract

The invention discloses a customer service system dialogue scheduling method and device, electronic equipment and a storage medium, and the method comprises the steps: responding to a user query request in a customer service system, and understanding a business scene of the user query request; according to the service scene of the query request of the user, judging whether a preset model hits the intention of the user or not; if not, dialogue scheduling is achieved in a preset mode. According to the invention, computing power resources are fully utilized, and intelligent dialogue scheduling of the customer service system is realized.

Description

Technical Field

[0001] The present application relates to the technical field of dialogue scheduling for a customer service system, and in particular to a dialogue scheduling method, device, electronic device, and storage medium for a customer service system. Background Art

[0002] With the development and rapid iteration of LLM, more and more intelligent services in the financial field have begun to apply LLM-related technologies.

[0003] How to use intelligent scheduling algorithms to fully utilize computing resources, reduce costs and increase efficiency while ensuring user experience. Summary of the invention

[0004] The embodiments of the present application provide a customer service system dialogue scheduling method, device, electronic device, and storage medium to implement the scheduling process involved in intelligent customer service interaction, thereby providing a service scheduling method with better performance.

[0005] The present application embodiment adopts the following technical solutions:

[0006] In a first aspect, an embodiment of the present application provides a customer service system dialogue scheduling method, wherein the dialogue scheduling method includes:

[0007] Respond to user query requests in the customer service system and understand the business scenarios of the user query requests;

[0008] According to the business scenario of the user query request, determine whether the preset model matches the user's intention;

[0009] If it is not hit, the dialogue scheduling is implemented in a pre-set manner.

[0010] In some embodiments, if the match is not found, the dialog scheduling is implemented in a preset manner, including:

[0011] If it is not matched, the service mode in the business scenario is selected for scheduling based on the computing service resource utilization rate and / or historical satisfaction data.

[0012] The service mode in the business scenario includes at least one of the following: manual, LLM model, and BOT model.

[0013] In some embodiments, the method further comprises:

[0014] According to the historical dissatisfaction scores in the historical satisfaction data, a manual service method in the business scenario is selected for scheduling;

[0015] According to the scores in the historical satisfaction data, the LLM model or BOT model service mode in the business scenario is selected for scheduling.

[0016] In some embodiments, the method further comprises:

[0017] According to the user intention hit rate and the computing power service resource utilization rate, the LLM model or BOT model service mode in the business scenario is selected for scheduling.

[0018] In some embodiments, the method further comprises:

[0019] According to the service resource utilization, select the LLM model or BOT model service mode in the business scenario for scheduling.

[0020] In some embodiments, the method further comprises:

[0021] In response to a user query request in the customer service system, query the high-frequency database to see if there are similar query requests;

[0022] If yes, then use the service behavior data in the high-frequency database;

[0023] By collecting user data, we can get the frequency of user query requests.

[0024] According to the frequency of the user query request questions, the service behavior data is regularly updated to the high-frequency database.

[0025] In some embodiments, the business scenario of solving the user query request includes:

[0026] Based on the LLM model, the business scenario to which the user query request belongs is calculated by combining the context information of the user query request.

[0027] In a second aspect, an embodiment of the present application further provides a customer service system dialogue scheduling device, wherein the dialogue scheduling device includes:

[0028] A response module, used to respond to user query requests in the customer service system and understand the business scenario of the user query request;

[0029] A judgment module, used to judge whether the preset model matches the user's intention according to the business scenario of the user's query request;

[0030] The scheduling module is used to implement the dialogue scheduling in a preset manner if a hit is not found.

[0031] In a third aspect, an embodiment of the present application further provides an electronic device, comprising: a processor; and a memory arranged to store computer executable instructions, wherein the executable instructions, when executed, cause the processor to perform the above method.

[0032] In a fourth aspect, an embodiment of the present application further provides a computer-readable storage medium, which stores one or more programs. When the one or more programs are executed by an electronic device including multiple application programs, the electronic device executes the above method.

[0033] At least one of the above technical solutions adopted in the embodiments of the present application can achieve the following beneficial effects: respond to user query requests in the customer service system and understand the business scenario of the user query request. Then, based on the business scenario of the user query request, determine whether the preset model matches the user's intention. If not, the dialogue scheduling is implemented in a pre-set manner. Through the above method, a service scheduling with better performance is provided, by modeling and processing the user's query request and historical behavior data, such as user satisfaction and historical interaction methods, combined with the current user's semantic request feedback. If the historical data is not met, the corresponding scheduling strategy is adopted to maximize the use of existing computing resources. It can not only improve the user's service experience and reduce costs and increase efficiency, but also improve the system's throughput and provide corresponding services to more users. BRIEF DESCRIPTION OF THE DRAWINGS

[0034] The drawings described herein are used to provide a further understanding of the present application and constitute a part of the present application. The illustrative embodiments of the present application and their descriptions are used to explain the present application and do not constitute an improper limitation on the present application. In the drawings:

[0035] Figure 1 A schematic diagram of the system architecture of the customer service system dialogue scheduling method in an embodiment of the present application;

[0036] Figure 2 This is a flowchart of a customer service system dialogue scheduling method in an embodiment of the present application;

[0037] Figure 3 This is a structural diagram of a customer service system dialogue scheduling device in an embodiment of the present application;

[0038] Figure 4 This is a schematic diagram of the structure of an electronic device in an embodiment of the present application. DETAILED DESCRIPTION

[0039] In order to make the purpose, technical solution and advantages of the present application clearer, the technical solution of the present application will be clearly and completely described below in combination with the specific embodiments of the present application and the corresponding drawings. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present application.

[0040] The technical solutions provided by various embodiments of the present application are described in detail below in conjunction with the accompanying drawings.

[0041] With the development and rapid iteration of LLM, more and more intelligent services in the financial field have begun to apply LLM-related technologies, which have the following shortcomings:

[0042] (1) There are more and more business scenarios that use dialogue to provide services, and the types of BOT models used are diverse. Taking mobile banking customer service as an example, it mainly involves multiple business lines such as financial management, credit, and credit cards. If the BOT model is used, there are both intelligent dialogue agents based on the LLM model and QA robots based on traditional knowledge bases.

[0043] (2) Regarding the cost of using the LLM large model, whether it is the hardware computing resources required for reply or the response speed, the existing LLM model service cannot compare with the small model dialogue system.

[0044] (3) The small model system has deficiencies in the accuracy of intent capture, contextual coherence, and response diversity.

[0045] At present, most of the implementation methods used by intelligent customer service are based on LLM agent mode. Through user-related knowledge base, documents and other data, RAG technology is used, combined with LLM's reasoning ability to achieve intelligent customer service response in the financial field. However, this solution has extremely high requirements on computing power. For example, one A100 GPU card only supports 6-8 customers concurrently. According to the existing business situation, only one application scenario of online customer service requires the support of 300+ cards, and the construction cost is extremely high. It cannot support the goal of rapid expansion and application of intelligent customer service in multiple business scenarios. At the same time, in real-time interaction scenarios, there are many cases where the same high-frequency questions are answered multiple times, which is a waste of computing power resources.

[0046] Finally, in terms of user services, since LLM needs to make inferences based on the above content, it uses long links to output tokens one by one, which has a certain impact on the system's concurrency and response time.

[0047] Therefore, how to use intelligent scheduling algorithms to fully utilize computing resources, reduce costs and increase efficiency while ensuring user experience is an urgent problem that needs to be solved.

[0048] In response to the above-mentioned shortcomings, the customer service system dialogue scheduling method in the embodiment of the present application uses user behavior data to provide recommendation services in response to the growing user needs and related application scenarios, as well as multiple BOT business scenarios. By adopting a flexible scheduling algorithm, the scheduling of intelligent dialogues is optimized, thereby achieving better response effects while consuming fewer resources.

[0049] like Figure 1 As shown, based on the user query request in the conversation context, it is input into the intelligent scheduling module. It is not only necessary to judge whether it belongs to the user's historical session data BOT high-frequency knowledge base data, but also to combine the user level, satisfaction evaluation, server resource utilization and other data for judgment. According to the business scenario, response service method, combined with server resources, user data and other information, it is allocated to the corresponding business scenario through the intelligent scheduling module, and combined with the corresponding processing method, the appropriate service provider is selected. In business scenario one, it includes LLM BOT1, traditional BOT and manual customer service. In business scenario two, it includes: LLMBOT2, QA BOT and manual customer service. In business scenario three, it includes LLM BOT3, BOT and manual customer service.

[0050] The present application embodiment provides a customer service system dialogue scheduling method, such as Figure 2 As shown, a flow chart of a method for scheduling a customer service system dialogue in an embodiment of the present application is provided, and the method at least includes the following steps S210 to S230:

[0051] Step S210, responding to a user query request in a customer service system, understanding a business scenario of the user query request.

[0052] According to the user query request in the customer service system, the business scenario of the user query request is judged and understood. In each business scenario, there may be a LLM model, a BOT model or a manual customer service path.

[0053] Step S220: Determine whether the preset model matches the user's intention based on the business scenario of the user's query request.

[0054] According to the business scenario of the user query request, it is further determined whether the preset model matches the user's intention. It can be understood that the "preset model" is the BOT model set in the customer service system. It is determined whether the BOT model has the possibility to complete the user's requested service. If yes, the BOT model is used directly.

[0055] Step S230: If no match is found, the dialogue scheduling is implemented in a preset manner.

[0056] If there is no match with the preset model, the dialogue scheduling will be implemented using the pre-set manual, BOT model or LLM model.

[0057] Through the above method, the scheduling and processing between multiple scenarios involving user interaction, such as LLM model customer service, BOT model customer service, and manual customer service, is realized. Due to different user preferences and habits, the services requested by query requests will be similar. Therefore, when a new query request arrives, the user's memory module (or module with memory function) is queried to see if there is a similar cache before, which serves as a reference for scheduling. If there are similar requests and the user service evaluation score is high, the corresponding path is directly scheduled.

[0058] When there is no similar reference method, the user's query is first identified by the business scenario, and the corresponding identification is generated for different business scenarios. Then, according to the identification, different scheduling methods are adopted through relevant calculation methods (such as satisfaction data, problem hits, server resource utilization, etc.).

[0059] Through the above method, by modeling and processing the user's query request and historical behavior data, such as user satisfaction and historical interaction mode, combined with the current user's semantic request, the corresponding scheduling strategy is adopted to maximize the use of existing computing resources. It can not only improve the user's service experience, reduce costs and increase efficiency, but also improve the system's throughput and provide corresponding services to more users.

[0060] Different from the related art, when combining the reasoning ability of the LLM model to realize the intelligent customer service response in the financial field, the system's concurrency and response time will be affected. Through the above method, in response to the user query request in the customer service system, the business scenario of the user query request is understood; according to the business scenario of the user query request, it is judged whether the preset model hits the user's intention; if it does not hit, the preset method is used to implement dialogue scheduling. In different business scenarios, by judging whether the preset model hits the user's intention, the best way to communicate with the user can be obtained more quickly, and when the preset model cannot meet the customer's query request, the preset method is promptly used to implement dialogue scheduling.

[0061] In one embodiment of the present application, if a hit is not found, a preset method is used to implement dialogue scheduling, including: if a hit is not found, a service method in the business scenario is selected for scheduling based on computing power service resource utilization and / or historical satisfaction data, and the service method in the business scenario includes at least one of the following: manual, LLM model, BOT model.

[0062] It can be understood that business scenarios include but are not limited to personal finance, financial management and other scenarios, and the manual, LLM model and BOT model configured in each scenario are different.

[0063] If it is determined that the preset model does not match the user's intention, the possibility of using the BOT model to complete the user's requested service is low. It is necessary to re-plan and schedule a service method suitable for the current scenario based on the computing power service resource utilization (server related) and historical satisfaction data (user habits / behavior / value related).

[0064] It can be understood that manual service means access to manual customer service. The LLM model, or Large Language Model, refers to a deep learning model trained with a large amount of text data that can generate natural language text or understand the meaning of language text. The core idea of ​​the LLM model is to learn the patterns and structures of natural language through large-scale unsupervised training, simulating the human language cognition and generation process. The BOT model refers to a small-scale QA customer service robot, etc.

[0065] In one embodiment of the present application, the method also includes: selecting a manual service mode in a business scenario for scheduling based on historical dissatisfaction scores in the historical satisfaction data; and selecting an LLM model or BOT model service mode in a business scenario for scheduling based on the scores in the historical satisfaction data.

[0066] Historical satisfaction data includes, but is not limited to, customer satisfaction ratings on services, percentage of manual transfers (Manual / Sum_interaction), customer business value (based on five levels), etc.

[0067] Taking the latest three months of historical data as an example, the closer the time is, the higher the time_weight is. As shown in the following table:

[0068]

[0069] as a satisfaction evaluation indicator.

[0070] Manual_Score = value*(Evaluation+exp(Manual / Sum_interaction)) / 4, as the manual score indicator, the higher the Manual_Score data, the more priority should be given to manual customer service. Similarly, according to the scores in the historical satisfaction data, the LLM model or BOT model service method in the business scenario can be selected for scheduling.

[0071] In one embodiment of the present application, the method further includes: selecting the LLM model or BOT model service mode in the business scenario for scheduling according to the user intention hit rate and the computing power service resource utilization rate.

[0072] User intent hit rate refers to the match between the preset model and user semantics:

[0073]

[0074] If the corresponding knowledge point or question can be accurately matched, the BOT model service method is used for scheduling. Otherwise, if it is not matched, the LLM model in the business scenario is considered for scheduling.

[0075] For example, comprehensive hit rate and resource utilization data:

[0076] Scores=Exact_match*Resource_score

[0077] Take the server resource utilization of 0.8 as an example:

[0078]

[0079] Scores are obtained by calculating the hit rate of user intentions and combining them with server resources, so as to select the appropriate BOT model service or LLM model service mode.

[0080] In one embodiment of the present application, the method further includes: selecting the LLM model or BOT model service mode in the business scenario for scheduling according to the service resource utilization rate.

[0081] Service resource utilization

[0082] Among them, level is the computing service resource utilization rate, which is between 0 and 1. The lower the service resource utilization rate, the more concurrent paths are available. In terms of scheduling, it is more suitable to respond with the LLM model that consumes more resources, which can more accurately understand the user's intention and provide accurate responses. Conversely, the higher the service resource utilization rate, the fewer concurrent paths are available, and it is more suitable to use the BOT model service method for scheduling.

[0083] In one embodiment of the present application, the method also includes: responding to a user query request in a customer service system, querying through a high-frequency database whether there is a similar query request; if so, using the service behavior data in the high-frequency database; obtaining the user query request problem frequency by collecting user data; and regularly updating the service behavior data to the high-frequency database based on the user query request problem frequency.

[0084] When the user completes the service request, the relevant data is collated according to the user's processing situation and incorporated into the user's high-frequency data. For example, the user's query is rewritten and the service behavior data (satisfaction, BOT model, knowledge base, etc.) is updated to facilitate the rapid dispatch of related services when a similar request occurs next time. In this way, the user's query request questions are updated according to the time period. The user's habits are recorded. In addition, infrequently used questions are deleted, and frequently used questions are retained, so that they are more matched than before.

[0085] Through the above method, when processing user data, the user's historical behavior data is also combined.

[0086] In one embodiment of the present application, solving the business scenario of the user query request includes: based on the LLM model, by combining the context information of the user query request, calculating the business scenario to which the user query request belongs.

[0087] For the user's query, the business scenario is first identified, and the corresponding identification is generated for different business scenarios. Then, according to the identification, different scheduling methods are adopted through relevant calculation methods (such as satisfaction data, problem hits, server resource utilization, etc.).

[0088] Regarding business scenarios, we can use a quantized and compressed model based on LLM to calculate the business scenario by combining the context information of the user query. Of course, other feasible models can also be used as long as they can meet the needs of business scenario understanding.

[0089] The embodiment of the present application also provides a customer service system dialogue scheduling device 300, such as Figure 3 As shown, a schematic diagram of the structure of the customer service system dialogue scheduling device in an embodiment of the present application is provided, and the customer service system dialogue scheduling device 300 at least includes: a response module 310, a judgment module 320, and a scheduling module 330, wherein:

[0090] In one embodiment of the present application, the response module 310 is specifically used to: respond to a user query request in a customer service system and understand a business scenario of the user query request.

[0091] According to the user query request in the customer service system, the business scenario of the user query request is judged and understood. Each business scenario may have an LLM model, a BOT model or a manual customer service path.

[0092] In one embodiment of the present application, the judgment module 320 is specifically used to: judge whether the preset model matches the user intention according to the business scenario of the user query request.

[0093] According to the business scenario of the user query request, it is further determined whether the preset model matches the user's intention. It can be understood that the "preset model" is the BOT model set in the customer service system. It is determined whether the BOT model has the possibility to complete the user's requested service. If yes, the BOT model is used directly.

[0094] In one embodiment of the present application, the scheduling module 330 is specifically used to: if there is no hit, implement the dialogue scheduling in a preset manner.

[0095] If there is no match with the preset model, the dialogue scheduling will be implemented using the pre-set manual, BOT model or LLM model.

[0096] In one embodiment of the present application, the scheduling module 330 is also used to

[0097] If it is not matched, the service mode in the business scenario is selected for scheduling based on the computing service resource utilization rate and / or historical satisfaction data.

[0098] The service mode in the business scenario includes at least one of the following: manual, LLM model, and BOT model.

[0099] In one embodiment of the present application, the scheduling module 330 is also used to

[0100] According to the historical dissatisfaction scores in the historical satisfaction data, a manual service method in the business scenario is selected for scheduling;

[0101] According to the scores in the historical satisfaction data, the LLM model or BOT model service mode in the business scenario is selected for scheduling.

[0102] In one embodiment of the present application, the scheduling module 330 is also used to

[0103] According to the user intention hit rate and the computing power service resource utilization rate, the LLM model or BOT model service mode in the business scenario is selected for scheduling.

[0104] In one embodiment of the present application, the scheduling module 330 is also used to

[0105] According to the service resource utilization, select the LLM model or BOT model service mode in the business scenario for scheduling.

[0106] In one embodiment of the present application, it also includes: an optimization module for

[0107] In response to a user query request in the customer service system, query the high-frequency database to see if there are similar query requests;

[0108] If yes, then use the service behavior data in the high-frequency database;

[0109] By collecting user data, we can get the frequency of user query requests.

[0110] According to the frequency of the user query request questions, the service behavior data is regularly updated to the high-frequency database.

[0111] In one embodiment of the present application, the response module 310 is also used to

[0112] Based on the LLM model, the business scenario to which the user query request belongs is calculated by combining the context information of the user query request.

[0113] It can be understood that the above-mentioned customer service system dialogue scheduling device can implement the various steps of the customer service system dialogue scheduling method provided in the aforementioned embodiment, and the relevant explanations about the customer service system dialogue scheduling method are applicable to the customer service system dialogue scheduling device, which will not be repeated here.

[0114] Figure 4 This is a schematic diagram of the structure of an electronic device according to an embodiment of the present application. Figure 4 At the hardware level, the electronic device includes a processor, and optionally also includes an internal bus, a network interface, and a memory. The memory may include a memory, such as a high-speed random access memory (RAM), and may also include a non-volatile memory (non-volatile memory), such as at least one disk storage. Of course, the electronic device may also include hardware required for other services.

[0115] The processor, network interface and memory can be interconnected through an internal bus, which can be an ISA (Industry Standard Architecture) bus, a PCI (Peripheral Component Interconnect) bus or an EISA (Extended Industry Standard Architecture) bus. The bus can be divided into an address bus, a data bus, a control bus, etc. For ease of representation, Figure 4 Only one bidirectional arrow is used in the diagram, but this does not mean that there is only one bus or only one type of bus.

[0116] The memory is used to store the program. Specifically, the program may include a program code, and the program code includes a computer operation instruction. The memory may include a memory and a non-volatile memory, and provides instructions and data to the processor.

[0117] The processor reads the corresponding computer program from the non-volatile memory into the memory and then runs it, forming a customer service system dialogue scheduling device at the logical level. The processor executes the program stored in the memory and is specifically used to perform the following operations:

[0118] Respond to user query requests in the customer service system and understand the business scenarios of the user query requests;

[0119] According to the business scenario of the user query request, determine whether the preset model matches the user's intention;

[0120] If it is not hit, the dialogue scheduling is implemented in a pre-set manner.

[0121] The above application Figure 2 The method performed by the customer service system dialogue scheduling device disclosed in the illustrated embodiment can be applied to a processor or implemented by a processor. The processor may be an integrated circuit chip with signal processing capabilities. In the implementation process, each step of the above method can be completed by an integrated logic circuit of hardware in the processor or an instruction in the form of software. The above processor can be a general-purpose processor, including a central processing unit (CPU), a network processor (NP), etc.; it can also be a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field-programmable gate array (FPGA) or other programmable logic devices, discrete gates or transistor logic devices, discrete hardware components. The methods, steps and logic block diagrams disclosed in the embodiments of the present application can be implemented or executed. The general-purpose processor can be a microprocessor or the processor can also be any conventional processor, etc. The steps of the method disclosed in the embodiments of the present application can be directly embodied as a hardware decoding processor for execution, or a combination of hardware and software modules in the decoding processor for execution. The software module can be located in a storage medium mature in the art such as a random access memory, a flash memory, a read-only memory, a programmable read-only memory, or an electrically erasable programmable memory, a register, etc. The storage medium is located in the memory, and the processor reads the information in the memory and completes the steps of the above method in combination with its hardware.

[0122] The electronic device may also perform Figure 2 A method for executing a customer service system dialogue scheduling device in the present invention, and realizing a customer service system dialogue scheduling device in the present invention Figure 2 The functions of the illustrated embodiment will not be described in detail in the embodiments of the present application.

[0123] The present application also provides a computer-readable storage medium, which stores one or more programs, wherein the one or more programs include instructions, which, when executed by an electronic device including multiple application programs, enable the electronic device to execute Figure 2 The method executed by the customer service system dialogue scheduling device in the embodiment shown is specifically used to execute:

[0124] Respond to user query requests in the customer service system and understand the business scenarios of the user query requests;

[0125] According to the business scenario of the user query request, determine whether the preset model matches the user's intention;

[0126] If it is not hit, the dialogue scheduling is implemented in a pre-set manner.

[0127] Those skilled in the art will appreciate that embodiments of the present invention may be provided as methods, systems, or computer program products. Therefore, the present invention may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Moreover, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0128] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0129] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0130] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0131] In a typical configuration, a computing device includes one or more processors (CPU), input / output interfaces, network interfaces, and memory.

[0132] The memory may include non-permanent storage in a computer-readable medium, random access memory (RAM) and / or non-volatile memory in the form of read-only memory (ROM) or flash RAM. The memory is an example of a computer-readable medium.

[0133] Computer readable media include permanent and non-permanent, removable and non-removable media that can be implemented by any method or technology to store information. Information can be computer readable instructions, data structures, program modules or other data. Examples of computer storage media include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technology, compact disk read-only memory (CD-ROM), digital versatile disk (DVD) or other optical storage, magnetic cassettes, magnetic tape magnetic disk storage or other magnetic storage devices or any other non-transmission media that can be used to store information that can be accessed by a computing device. As defined herein, computer readable media does not include temporary computer readable media (transitory media), such as modulated data signals and carrier waves.

[0134] It should also be noted that the terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, commodity or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, commodity or device. In the absence of more restrictions, the elements defined by the sentence "comprises a ..." do not exclude the existence of other identical elements in the process, method, commodity or device including the elements.

[0135] Those skilled in the art will appreciate that the embodiments of the present application may be provided as methods, systems or computer program products. Therefore, the present application may adopt the form of a complete hardware embodiment, a complete software embodiment or an embodiment in combination with software and hardware. Moreover, the present application may adopt the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) that contain computer-usable program code.

[0136] The above is only an embodiment of the present application and is not intended to limit the present application. For those skilled in the art, the present application may have various changes and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application should be included in the scope of the claims of the present application.

Claims

1. A customer service system dialogue scheduling method, wherein: The dialogue scheduling method comprises: Respond to user query requests in the customer service system and understand the business scenarios of the user query requests; According to the business scenario of the user query request, determine whether the preset model matches the user's intention; If it is not hit, the dialogue scheduling is implemented in a pre-set manner.

2. The method of claim 1, wherein: If the match is not found, the dialog scheduling is implemented in a preset manner, including: If it is not matched, the service mode in the business scenario is selected for scheduling based on the computing service resource utilization rate and / or historical satisfaction data. The service mode in the business scenario includes at least one of the following: manual, LLM model, and BOT model.

3. The method according to claim 2, further comprising: According to the historical dissatisfaction scores in the historical satisfaction data, a manual service method in the business scenario is selected for scheduling; According to the scores in the historical satisfaction data, the LLM model or BOT model service mode in the business scenario is selected for scheduling.

4. The method according to claim 2, further comprising: According to the user intention hit rate and the computing power service resource utilization rate, the LLM model or BOT model service mode in the business scenario is selected for scheduling.

5. The method according to claim 2, further comprising: According to the service resource utilization, select the LLM model or BOT model service mode in the business scenario for scheduling.

6. The method according to claim 1, further comprising: In response to a user query request in the customer service system, query the high-frequency database to see if there are similar query requests; If yes, then use the service behavior data in the high-frequency database; By collecting user data, we can get the frequency of user query requests. According to the frequency of the user query request questions, the service behavior data is regularly updated to the high-frequency database.

7. The method of claim 1, wherein: The business scenario of solving the user query request includes: Based on the LLM model, the business scenario to which the user query request belongs is calculated by combining the context information of the user query request.

8. A customer service system dialogue scheduling device, wherein: The dialogue scheduling device comprises: A response module, used to respond to user query requests in the customer service system and understand the business scenario of the user query request; A judgment module, used to judge whether the preset model matches the user's intention according to the business scenario of the user's query request; The scheduling module is used to implement the dialogue scheduling in a preset manner if a hit is not found.

9. An electronic device, comprising: processor; as well as A memory arranged to store computer executable instructions, which when executed cause the processor to perform the method of any one of claims 1 to 7.

10. A computer-readable storage medium storing one or more programs, which, when executed by an electronic device including a plurality of application programs, causes the electronic device to execute any one of the methods of claims 1 to 7.

Images