Dialogue quality inspection method, computing device and storage medium

Abstract

The invention discloses a dialogue quality inspection method. The dialogue quality inspection method comprises the steps of obtaining dialogue data between a server and a serviced person and a target quality inspection item corresponding to the dialogue data; based on the target quality inspection item, obtaining a corresponding dialogue content extraction rule and a first prompt word; according to a dialogue content extraction rule, dialogue content to be subjected to quality inspection is extracted from the dialogue data; and performing quality inspection on the dialogue content to be subjected to quality inspection based on the first prompt word to generate a dialogue quality inspection result According to the dialogue quality inspection method, the computing device and the storage medium provided by the invention, the quality inspection efficiency, the quality inspection accuracy, the quality inspection fineness and the quality inspection comprehensiveness of the dialogue content of the server can be remarkably improved.

Description

Technical Field

[0001] This application relates to the field of artificial intelligence technology. Specifically, it relates to a dialogue quality inspection method, a computing device, and a storage medium. Background Art

[0002] With the development of artificial intelligence technology, intelligent agent process orchestration has gradually become an important means for various industries to improve the level of automation. Intelligent agent process orchestration is a technology that manages workflows through distributed autonomous agents, combining artificial intelligence algorithms, machine learning models, and decision-making logic to improve the execution efficiency and automation level of tasks. Currently, this technology has been widely applied in fields such as enterprise customer service automation, industrial manufacturing, logistics management, and fintech.

[0003] In intelligent customer service systems, the quality inspection of customer service dialogue content is an important link to improve the customer service experience. Traditional quality inspection methods rely on manual review or rule-based automatic detection methods, which often have problems such as long time consumption, limited coverage, and lack of flexibility. In recent years, with the introduction of large language model (LLM) technology, intelligent customer service systems have made significant progress in the automation and accuracy of quality inspection. However, despite this, existing solutions still face many problems in practical applications.

[0004] Existing solutions usually rely on large language models to perform integrated quality inspection and analysis on the overall dialogue text, resulting in low quality inspection efficiency and inaccurate quality inspection results for such quality inspection solutions. Summary of the Invention

[0005] To solve the above problems, the present invention proposes a dialogue quality inspection method, a computing device, and a storage medium, which can significantly improve the quality inspection efficiency, accuracy, fineness, and comprehensiveness of dialogue content.

[0006] The present invention provides a dialogue quality inspection method, the method comprising: obtaining dialogue data between a service provider and a service recipient and target quality inspection items corresponding to the dialogue data; based on the target quality inspection items, obtaining corresponding dialogue content extraction rules and a first prompt word; extracting dialogue content to be quality inspected from the dialogue data according to the dialogue content extraction rules; and performing quality inspection on the dialogue content to be quality inspected based on the first prompt word to generate a dialogue quality inspection result.

[0007] In one embodiment, the step of obtaining corresponding dialogue content extraction rules and a first prompt word based on the target quality inspection items comprises: querying the dialogue content extraction rules and the first prompt word respectively corresponding to different quality inspection items in a quality inspection item business table, and obtaining the dialogue content extraction rules and the first prompt word corresponding to the target quality inspection items.

[0008] In one embodiment, the step of extracting the conversation content to be quality-checked from the conversation data according to the conversation content extraction rule includes: analyzing the conversation data to identify the conversation stage or conversation structure of the conversation data; determining the conversation content of the conversation stage corresponding to the conversation content extraction rule or the conversation content of the corresponding conversation structure as the conversation content to be quality-checked.

[0009] In one embodiment, the conversation content to be quality-checked includes at least one of the following: the service provider's conversation content in the conversation data; the service provider's conversation content and the customer's conversation content in the conversation data.

[0010] In one embodiment, the step of performing quality check on the conversation content to be quality-checked based on the first prompt word and generating a conversation quality check result includes: when the conversation content to be quality-checked only includes the service provider's conversation content, identifying whether the service provider's conversation content meets the conditions set by the first prompt word; if it meets, determining that the conversation content to be quality-checked passes the quality check of the target quality check item; if it does not meet, determining that the conversation content to be quality-checked fails the quality check of the target quality check item.

[0011] In one embodiment, the step of performing quality check on the conversation content to be quality-checked based on the first prompt word and generating a conversation quality check result includes: when the conversation content to be quality-checked includes the service provider's conversation content and the customer's conversation content, identifying the service provider's conversation content in the conversation content to be quality-checked based on the first prompt word to obtain a target identification result; querying the second prompt words corresponding to different identification results in the first prompt word identification result service table to obtain the second prompt word corresponding to the target identification result; based on the second prompt word corresponding to the identification result, identifying whether the service provider's conversation content in the conversation content to be quality-checked meets the conditions set by the second prompt word; if it meets, determining that the conversation content to be quality-checked passes the quality check of the target quality check item; if it does not meet, determining that the conversation content to be quality-checked fails the quality check of the target quality check item.

[0012] In one embodiment, the step of extracting the conversation content to be quality-checked from the conversation data according to the conversation content extraction rule includes: when the conversation content extraction rule includes extracting historical conversation data, obtaining the historical conversation data associated with the conversation data; obtaining the conversation content to be quality-checked from the conversation data and the historical conversation data according to the conversation content extraction rule.

[0013] In one embodiment, the step of performing quality inspection on the conversation content to be quality-inspected based on the first prompt word and generating a conversation quality inspection result includes: determining a target quality inspection model based on the target quality inspection item; and invoking the target quality inspection model to perform quality inspection on the conversation content to be quality-inspected based on the first prompt word and generate a conversation quality inspection result.

[0014] The present invention also provides a computing device, including a processor and a memory storing a computer program. When the processor runs the computer program, the steps of the conversation quality inspection method as described above are implemented.

[0015] The present invention also provides a storage medium storing a computer program. When the computer program is executed by a processor, the steps of the conversation quality inspection method as described above are implemented.

[0016] The conversation quality inspection method, computing device, and storage medium provided by the present invention obtain conversation data between a service provider and a service recipient and the target quality inspection item corresponding to the conversation data. Based on the target quality inspection item, the corresponding conversation content extraction rule and the first prompt word are obtained. According to the conversation content extraction rule, the conversation content to be quality-inspected is extracted from the conversation data. Furthermore, quality inspection is performed on the conversation content to be quality-inspected based on the first prompt word to generate a conversation quality inspection result, which can significantly improve the quality inspection efficiency, accuracy, fineness, and comprehensiveness of the service provider's conversation content. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] Figure 1 It is a flowchart of the conversation quality inspection method according to an embodiment of the present invention.

[0018] Figure 2 is Figure 1 a flowchart of step S14 in

[0019] Figure 3 It is a structural diagram of the computing device according to an embodiment of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0020] The foregoing and other technical contents, features, and effects of the present invention will be clearly presented in the following detailed description of the preferred embodiments with reference to the drawings. Through the description of the specific embodiments, a more in-depth and specific understanding of the technical means and effects adopted by the present invention to achieve the predetermined purpose can be obtained. However, the drawings are only for reference and illustration, and are not used to limit the present invention.

[0021] To further elaborate on the technical means and effects adopted by the present invention to achieve the predetermined invention purpose, the present invention is described in detail below with reference to the drawings and preferred embodiments.

[0022] Figure 1Flow schematic diagram of the dialogue quality inspection method according to an embodiment of the present invention.

[0023] As Figure 1 shown, the dialogue quality inspection method provided in this embodiment includes the following steps:

[0024] Step S11: Obtain the dialogue data between the service provider and the service recipient and the target quality inspection items corresponding to the dialogue data.

[0025] Specifically, the dialogue data is the data generated during the call and / or communication between the service provider and the service recipient. The dialogue data includes the dialogue content of the service recipient and the dialogue content of the service provider. The dialogue data is in text format and / or audio format. When obtaining the dialogue data, the corresponding quality inspection task of the dialogue data will also be obtained. The quality inspection task records which quality inspection needs to be performed on the dialogue data, so as to determine the target quality inspection items corresponding to the dialogue data. The dialogue data can correspond to a single or multiple target quality inspection items.

[0026] Preferably, in this embodiment, the quality inspection items include at least one of detecting whether the service provider uses a standard opening statement, detecting whether the service provider uses a standard closing statement, detecting whether the service provider guides the service recipient to conduct a satisfaction evaluation before hanging up, detecting whether the service provider ensures that the service recipient has no other business needs before hanging up, detecting whether the service provider uses polite language, detecting whether the service provider uses prohibited service language, detecting the accuracy of the service provider's identification of the service recipient's needs, detecting the standardization of the service provider's response to the service recipient's questions, detecting the personalized matching degree of the service provider's response to the service recipient's questions, detecting the correlation between the emotional changes of the service recipient and the service provider's response, detecting whether the emotion of the service recipient deteriorates, and detecting the coherence of the semantic logic of the service provider.

[0027] Step S12: Based on the target quality inspection items, obtain the corresponding dialogue content extraction rules and the first prompt words.

[0028] Specifically, query the dialogue content extraction rules and the first prompt words corresponding to different quality inspection items in the quality inspection item business table, and obtain the dialogue content extraction rules and the first prompt words corresponding to the target quality inspection items.

[0029] Among them, the quality inspection item business table records the dialogue content extraction rules and the first prompt words corresponding to each quality inspection item; the dialogue content extraction rules corresponding to the target quality inspection items are used to screen out the dialogue content related to the target quality inspection items from the dialogue data, and it determines what kind of dialogue content needs to be extracted as the data to be quality inspected; the first prompt words are used to guide the quality inspection model (such as an AI model) to identify and / or detect the dialogue content to be quality inspected, so that the quality inspection model outputs the identification and / or detection results.

[0030] In one embodiment, the quality inspection item service form is updated based on the received adjustment information of the quality inspection item service form. That is, the present invention can dynamically update the quality inspection items, the dialogue content extraction rules, and the prompt words, which can ensure that the system is applicable to different business requirements and improve the adjustability and adaptability of the quality inspection model.

[0031] In one example, the quality inspection item service form is as follows:

[0032]

[0033]

[0034] Step S13: Extract the dialogue content to be quality inspected from the dialogue data according to the dialogue content extraction rules.

[0035] Specifically, analyze the dialogue data to identify the dialogue stage or dialogue structure of the dialogue data; determine the dialogue content of the dialogue stage corresponding to the dialogue content extraction rules, or the dialogue content of the corresponding dialogue structure, as the dialogue content to be quality inspected.

[0036] Specifically, based on the call content, the dialogue data is divided into multiple dialogue stages (such as the opening guidance stage, the question and answer stage, etc.), and / or based on the dialogue structure, the dialogue data is divided into different structures (such as the first dialogue structure including all the dialogue content of the service provider, the second dialogue structure including all the dialogue content of the service recipient, etc.). According to the dialogue content extraction rules, the corresponding dialogue structure and / or dialogue stage is set as the dialogue content to be quality inspected. The dialogue content to be quality inspected is used to provide the content for the quality inspection model to identify and / or detect.

[0037] Specifically, the extracted dialogue content to be quality inspected only includes the dialogue content of the service provider in the dialogue data, or includes the dialogue content of the service provider and the service recipient in the dialogue data, or only includes the dialogue content of the service provider in the dialogue data.

[0038] In one embodiment, when the quality inspection item is to detect the personalized matching degree of the service provider's reply to the service recipient's question, according to the dialogue content extraction rules, it will not only extract the dialogue content of the service provider and the service recipient in the question and answer stage in the dialogue data, but also obtain the historical dialogue data associated with the dialogue data, extract the dialogue content of the service provider and the service recipient in the question and answer stage in the associated historical dialogue data, and set the dialogue content of the service provider and the service recipient in the question and answer stage in the extracted dialogue data, and the dialogue content of the service provider and the service recipient in the question and answer stage in the associated historical dialogue data as the dialogue content to be quality inspected. This can enable the present invention to identify whether the service provider provides a corresponding personalized question solution according to the historical dialogue data of the service recipient when replying to the service recipient's question.

[0039] The present invention splits the dialogue data, confines the quality inspection content of different quality inspection items within a smaller range, enables the prompt words to be more targeted at specific quality inspection items, reduces the complexity of the quality inspection task, and ensures that the output of the quality inspection result is more accurate and meets the expectations. The miniaturized splitting also reduces the amount of data processed by the quality inspection model at one time during quality inspection, effectively improves the parsing and response speed of the quality inspection model, and realizes a smooth experience at the real-time dialogue level.

[0040] Step S14: Perform quality inspection on the dialogue content to be quality inspected based on the first prompt word, and generate a dialogue quality inspection result.

[0041] As Figure 2 shown, step S14 includes:

[0042] Step S141: Identify whether the dialogue content to be quality inspected only includes the dialogue content of the service provider.

[0043] Specifically, when it is identified that the dialogue content to be quality inspected only includes the dialogue content of the service provider, go to step S142: Identify whether the dialogue content of the service provider meets the conditions set by the first prompt word; when it is identified that the dialogue content to be quality inspected does not only include the dialogue content of the service provider, go to step S145: Identify the dialogue content of the service recipient in the dialogue content to be quality inspected based on the first prompt word, and obtain the target recognition result.

[0044] Specifically, in step S142, if, based on the first prompt word, it is identified that the dialogue content of the service provider in the dialogue content to be quality inspected meets the conditions set by the first prompt word, go to step S143: Determine that the dialogue content to be quality inspected passes the quality inspection of the target quality inspection item; if, based on the first prompt word, it is identified that the dialogue content of the service provider in the dialogue content to be quality inspected does not meet the conditions set by the first prompt word, go to step S144: Determine that the dialogue content to be quality inspected fails the quality inspection of the target quality inspection item. For example, when the target quality inspection item is to detect whether the service provider uses prohibited words, based on the corresponding first prompt word, identify whether the dialogue content of the service provider in the dialogue data does not include any of the prohibited words 1, prohibited words 2... prohibited words N. When it is identified that the dialogue content of the service provider does not include any prohibited words, determine that the dialogue content of the service provider meets the conditions set by the first prompt word, and go to step S143. When it is identified that the dialogue content of the service provider includes at least one prohibited word, determine that the dialogue content of the service provider does not meet the conditions set by the first prompt word, and go to step S144.

[0045] Specifically, during quality inspection, according to the corresponding target quality inspection item, determine the target quality inspection model corresponding to the target quality inspection item, and call the target quality inspection model to use the first prompt word corresponding to the target quality inspection item to perform quality inspection on the dialogue content to be quality inspected corresponding to the target quality inspection item, so as to generate the dialogue quality inspection result of the target quality inspection item. Different quality inspection models record the first prompt word recognition result business table corresponding to the corresponding quality inspection item. The first prompt word recognition result business table records the second prompt words corresponding to the recognition results of different first prompt words. The second prompt word is used to guide the quality inspection model to further detect the dialogue content to be quality inspected, so that the quality inspection model outputs the detection result.

[0046] Specifically, in step S145, based on the first prompt word, identify the dialogue content of the served party in the dialogue content to be quality inspected. After obtaining the target recognition result, query the corresponding first prompt word recognition result business table to obtain the second prompt word corresponding to the target recognition result. When the second prompt word corresponding to the target recognition result is not obtained, generate an error message. For example, when the target quality inspection item is to detect the standard degree of the service provider's reply to the served party's question, based on the corresponding first prompt word, when the question of the served party in the dialogue content of the served party is identified as question A, according to the second prompt words corresponding to the recognition results of different first prompt words recorded in the first prompt word recognition result business table corresponding to the target quality inspection item, determine the second prompt word corresponding to question A (i.e., the target recognition result); when the standard solutions to question A include A1, A2, and A3, the second prompt word recorded for question A is to identify whether the dialogue content of the service provider includes at least one of all the key words of A1 or all the key words of A2 or all the key words of A3.

[0047] Step S146: Based on the second prompt word corresponding to the target recognition result, identify whether the dialogue content of the service provider in the dialogue content to be quality inspected meets the conditions set by the second prompt word.

[0048] Specifically, if, based on the second prompt word, it is recognized that the service provider's conversation content in the conversation content to be quality inspected meets the conditions set by the second prompt word, then step S143 is entered; if, based on the second prompt word, it is recognized that the service provider's conversation content in the conversation content to be quality inspected does not meet the conditions set by the second prompt word, then step S144 is entered. For example, when the quality inspection item is to detect the standard degree of the service provider's response to the service recipient's question, when the corresponding second prompt word is obtained according to the recognition result of the first prompt word, based on the corresponding second prompt word, it is recognized whether the service provider's conversation content includes at least one of all the keywords of A1 or all the keywords of A2 or all the keywords of A3. When the service provider's conversation content includes at least one of all the keywords of A1 or all the keywords of A2 or all the keywords of A3, step S143 is entered; when the service provider's conversation content does not include all the keywords of A1 and / or all the keywords of A2 and / or all the keywords of A3, step S144 is entered.

[0049] In one embodiment, before step S145, it includes recognizing whether the conversation content to be quality inspected only includes the service recipient's conversation content. If not, then step S145 is entered; if so, then based on the first prompt word, it is recognized whether the service recipient's conversation content in the conversation content to be quality inspected meets the conditions set by the first prompt word. When the service recipient's conversation content meets the conditions set by the first prompt word, step S143 is entered; when the service recipient's conversation content does not meet the conditions set by the first prompt word, step S144 is entered. For example, when the quality inspection item is to detect whether the service recipient's mood deteriorates, the conversation content extraction rule is to extract the second conversation structure. The conversation content to be quality inspected obtained according to this conversation content extraction rule only includes the service recipient's conversation content. Based on the corresponding first prompt word, it is recognized the emotional state of the conversation content in the conversation content to be quality inspected at multiple conversation stages, and it is recognized whether the emotional state of the conversation content does not decrease as the conversation stage progresses. When the emotional state of the conversation content does not decrease as the conversation stage progresses, step S143 is entered; when the emotional state of the conversation content decreases as the conversation stage progresses, step S144 is entered. This emotional analysis mechanism not only improves the comprehensiveness of quality inspection, but also helps enterprises optimize the service provider strategy, reduce the service recipient churn rate, and improve the service recipient experience.

[0050] When the conversation content to be quality inspected of the present invention includes both the service recipient's conversation content and the service provider's conversation content, it can further match the corresponding second prompt word according to the question raised by the service recipient, and accordingly check the standard degree of the service provider's response to this question. Through this double-layer prompt word quality inspection mechanism, the present invention can effectively improve the accuracy and applicability of quality inspection, enabling the quality inspection model to intelligently identify the service quality of service providers in different scenarios.

[0051] Specifically, during quality inspection, the quality inspection model identifies the conversation content to be inspected through speech recognition and / or text recognition.

[0052] In one embodiment, there are multiple quality inspection models. Each quality inspection item is only inspected using the corresponding quality inspection model. Each quality inspection model corresponds to only one quality inspection item, and each quality inspection item has a dedicated quality inspection model.

[0053] The present invention can hand over multiple target quality inspection items corresponding to the conversation data to the corresponding quality inspection models for parallel processing. It can not only improve the quality inspection efficiency, but also improve the training effect of different quality inspection models, making the analysis of the quality inspection model more targeted and professional.

[0054] In one embodiment, after all the target quality inspection items corresponding to the conversation data are completed, a final quality inspection score and / or quality inspection rating is generated based on the quality inspection results of each quality inspection item of the conversation data.

[0055] In summary, the conversation quality inspection method provided by the present invention obtains the conversation data between the service provider and the service recipient and the target quality inspection items corresponding to the conversation data. Based on the target quality inspection items, the corresponding conversation content extraction rules and the first prompt words are obtained. According to the conversation content extraction rules, the conversation content to be inspected is extracted from the conversation data, and then the conversation content to be inspected is inspected based on the first prompt words to generate a conversation quality inspection result, which can significantly improve the quality inspection efficiency, accuracy, fineness and comprehensiveness of the service provider's conversation content.

[0056] Based on the same inventive concept as the foregoing embodiments, an embodiment of the present invention provides a computing device, as Figure 3 shown. The computing device includes: a processor 310 and a memory 311 storing a computer program; wherein, Figure 3 The processor 310 shown in does not refer to the number of processors 310 being one, but only refers to the positional relationship of the processor 310 relative to other components. In actual applications, the number of processors 310 can be one or more; similarly, Figure 3 The memory 311 shown in also has the same meaning, that is, it only refers to the positional relationship of the memory 311 relative to other components. In actual applications, the number of memories 311 can be one or more. When the processor 310 runs the computer program, the above-mentioned conversation quality inspection method is implemented.

[0057] The computing device may further include: at least one network interface 312. Each component in the computer device is coupled together through a bus system 313. It can be understood that the bus system 313 is used to realize the connection and communication between these components. The bus system 313 includes not only a data bus, but also a power bus, a control bus and a status signal bus. However, for the sake of clarity, in Figure 3Label various buses as bus system 313.

[0058] Among them, the memory 311 can be a volatile memory or a non-volatile memory, or can include both volatile and non-volatile memories. Among them, the non-volatile memory can be a read-only memory (ROM, Read Only Memory), a programmable read-only memory (PROM, Programmable Read-Only Memory), an erasable programmable read-only memory (EPROM, Erasable Programmable Read-Only Memory), an electrically erasable programmable read-only memory (EEPROM, Electrically Erasable Programmable Read-Only Memory), a ferromagnetic random access memory (FRAM, ferromagnetic random access memory), a flash memory (Flash Memory), a magnetic surface memory, an optical disc, or a compact disc read-only memory (CD-ROM, Compact Disc Read-Only Memory); the magnetic surface memory can be a disk memory or a tape memory. The volatile memory can be a random access memory (RAM, Random Access Memory), which is used as an external cache. By way of example but not limitation, many forms of RAM are available, such as a static random access memory (SRAM, Static Random Access Memory), a synchronous static random access memory (SSRAM, Synchronous Static Random Access Memory), a dynamic random access memory (DRAM, Dynamic Random Access Memory), a synchronous dynamic random access memory (SDRAM, Synchronous Dynamic Random Access Memory), a double data rate synchronous dynamic random access memory (DDR SDRAM, Double Data Rate Synchronous Dynamic Random Access Memory), an enhanced synchronous dynamic random access memory (ESDRAM, Enhanced Synchronous Dynamic Random Access Memory), a sync link dynamic random access memory (SLDRAM, SyncLink Dynamic Random Access Memory), a direct rambus random access memory (DRRAM, Direct Rambus Random Access Memory).The memory 311 described in the embodiments of the present invention is intended to include, but is not limited to, these and any other suitable types of memory.

[0059] The memory 311 in the embodiments of the present invention is used to store various types of data to support the operation of the computing device. Examples of such data include: any computer programs for operating on the computing device, such as an operating system and application programs. Among them, the operating system includes various system programs, such as a framework layer, a core library layer, a driver layer, etc., for implementing various basic services and processing hardware-based tasks. Application programs can include various application programs, such as a Media Player, a Browser, etc., for implementing various application services. Here, the program for implementing the method of the embodiments of the present invention can be included in the application program.

[0060] Based on the same inventive concept as the foregoing embodiments, this embodiment also provides a storage medium in which a computer program is stored. The storage medium can be a ferromagnetic random access memory (FRAM), a read-only memory (ROM), a programmable read-only memory (PROM), an erasable programmable read-only memory (EPROM), an electrically erasable programmable read-only memory (EEPROM), a flash memory, a magnetic surface memory, an optical disc, or a compact disc read-only memory (CD-ROM), etc.; it can also be various devices including one or any combination of the above memories, such as a mobile phone, a computer, a tablet device, a personal digital assistant, etc. When the computer program stored in the storage medium is run by a processor, the above-mentioned dialogue quality inspection method is implemented. For the specific step flow implemented when the computer program is executed by the processor, please refer to Figure 1 、 Figure 2 the description of the illustrated embodiment, which will not be repeated here.

[0061] The dialogue quality inspection method, computing device, and storage medium provided by the present invention obtain the dialogue data between the service provider and the service recipient and the target quality inspection items corresponding to the dialogue data. Based on the target quality inspection items, the corresponding dialogue content extraction rules and the first prompt words are obtained. According to the dialogue content extraction rules, the dialogue content to be quality inspected is extracted from the dialogue data. Furthermore, the dialogue content to be quality inspected is quality inspected based on the first prompt words to generate a dialogue quality inspection result, which can significantly improve the quality inspection efficiency, accuracy, fineness, and comprehensiveness of the service provider's dialogue content.

[0062] It should be understood that although the steps in the flowchart of the accompanying drawings are shown in sequence according to the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless otherwise clearly stated in this article, the execution of these steps has no strict order limit, and they can be executed in other orders. Moreover, at least a part of the steps in the accompanying drawings may include multiple sub-steps or multiple stages. These sub-steps or stages are not necessarily executed at the same time, but can be executed at different times, and their execution order is not necessarily sequential, but can be executed alternately or alternately with at least a part of other steps or sub-steps or stages of other steps.

[0063] Through the description of the above embodiments, those skilled in the art can clearly understand that the embodiments of the present invention can be implemented by means of software plus a necessary general hardware platform. Based on such an understanding, the technical solutions of the embodiments of the present invention can be embodied in the form of a software product, which can be stored in a non-volatile storage medium (which can be a CD-ROM, USB flash drive, mobile hard disk, etc.), including several instructions for causing a computer device (which can be a personal computer, server, or network device, etc.) to execute the methods of the various implementation scenarios of the embodiments of the present invention.

[0064] The above are only the preferred embodiments of the present invention and do not impose any form of limitation on the present invention. Although the present invention has been disclosed above with the preferred embodiments, it is not intended to limit the present invention. Any person skilled in the art can make some changes or modifications to the above-disclosed technical content to be equivalent embodiments within the scope of the technical solutions of the present invention. However, as long as it does not depart from the content of the technical solutions of the present invention, any simple modification, equivalent change, and modification made to the above embodiments based on the technical essence of the present invention still fall within the scope of the technical solutions of the present invention.

Claims

1. A method for dialogue quality inspection, characterized in that, The method includes: Obtaining the conversation data between the service provider and the service recipient and the target quality inspection items corresponding to the conversation data; Based on the target quality inspection items, obtaining the corresponding conversation content extraction rules and the first prompt words; According to the conversation content extraction rules, extracting the conversation content to be quality inspected from the conversation data; Based on the first prompt words, performing quality inspection on the conversation content to be quality inspected to generate a conversation quality inspection result.

2. The dialogue quality inspection method according to claim 1, wherein The step of obtaining the corresponding conversation content extraction rules and the first prompt words based on the target quality inspection items includes: Querying the conversation content extraction rules and the first prompt words respectively corresponding to different quality inspection items in the quality inspection item service table, and obtaining the conversation content extraction rules and the first prompt words corresponding to the target quality inspection items.

3. The dialogue quality inspection method according to claim 1, characterized in that, The step of extracting the conversation content to be quality inspected from the conversation data according to the conversation content extraction rules includes: Analyzing the conversation data to identify the conversation stage or conversation structure of the conversation data; Determining the conversation content of the conversation stage corresponding to the conversation content extraction rules or the conversation content of the corresponding conversation structure as the conversation content to be quality inspected.

4. The dialogue quality inspection method according to claim 3, wherein The conversation content to be quality inspected includes at least one of the following: the conversation content of the service provider in the conversation data; The conversation content of the service provider and the conversation content of the service recipient in the conversation data.

5. The dialogue quality inspection method according to claim 4, characterized in that, The step of performing quality inspection on the conversation content to be quality inspected based on the first prompt words to generate a conversation quality inspection result includes: When the conversation content to be quality inspected only includes the conversation content of the service provider, identifying whether the conversation content of the service provider meets the conditions set by the first prompt words; If it meets, determining that the conversation content to be quality inspected passes the quality inspection of the target quality inspection item; If it does not meet, determining that the conversation content to be quality inspected fails the quality inspection of the target quality inspection item.

6. The dialogue quality inspection method according to claim 4, wherein The step of performing quality inspection on the conversation content to be quality inspected based on the first prompt words to generate a conversation quality inspection result includes: When the conversation content to be quality inspected includes the conversation content of the service provider and the conversation content of the service recipient, identifying the conversation content of the service recipient in the conversation content to be quality inspected based on the first prompt words to obtain a target recognition result; Querying the second prompt words respectively corresponding to different recognition results in the first prompt word recognition result service table, and obtaining the second prompt words corresponding to the target recognition result; Identifying whether the conversation content of the service provider in the conversation content to be quality inspected meets the conditions set by the second prompt words; If it meets, determining that the conversation content to be quality inspected passes the quality inspection of the target quality inspection item; If it does not meet, determining that the conversation content to be quality inspected fails the quality inspection of the target quality inspection item.

7. The dialogue quality inspection method according to claim 3, characterized in that The step of extracting the conversation content to be quality inspected from the conversation data according to the conversation content extraction rules includes: When the conversation content extraction rules include extracting historical conversation data, obtaining the historical conversation data associated with the conversation data; According to the conversation content extraction rules, obtaining the conversation content to be quality inspected from the conversation data and the historical conversation data.

8. The dialogue quality inspection method according to claim 1, characterized in that, The step of performing quality inspection on the conversation content to be quality-inspected based on the first prompt word and generating a conversation quality inspection result includes: Determine a target quality inspection model based on the target quality inspection item; Call the target quality inspection model to perform quality inspection on the conversation content to be quality-inspected based on the first prompt word and generate a conversation quality inspection result.

9. A computing device, characterized in that, It includes a processor and a memory storing a computer program. When the processor runs the computer program, the steps of the conversation quality inspection method described in any one of claims 1 to 8 are implemented.

10. A storage medium storing a computer program, characterized in that, When the computer program is executed by the processor, the steps of the conversation quality inspection method described in any one of claims 1 to 8 are implemented.

Images