A work order quality inspection method, device, equipment and readable storage medium

Abstract

The application discloses a work order quality inspection method and device, equipment and a readable storage medium. The method comprises the following steps: obtaining conversation information of an agent and a customer in a work order; identifying a customer intention according to the conversation information; determining a quality inspection item combination matched with the customer intention, wherein the quality inspection item combination comprises a plurality of quality inspection items; scoring each quality inspection item in the quality inspection item combination according to the conversation information and a preset quality inspection item scoring standard; summarizing the scoring results of each quality inspection item in the quality inspection item combination to obtain a total quality inspection score; and determining a quality inspection result according to the total quality inspection score. The application can obtain the conversation information of the agent and the customer in the work order, use a server to detect each quality inspection item of the work order, and determine the quality inspection result of the work order. In this process, the server can complete the quality inspection process of the work order, manual conversation playback for quality inspection is not needed, manual operation for quality inspection can be reduced, and the quality inspection efficiency is improved.

Description

Technical Field

[0001] This application relates to the field of computer intelligence, and more specifically, to a work order quality inspection method, apparatus, device, and readable storage medium. Background Technology

[0002] With the development of the internet economy and e-commerce platforms, customer service has emerged. Customer service is an important way for call center systems to provide services to customers. It refers to the process by which agents provide services to customers through the call center's support system. When customers call to inquire about pre-sales or after-sales issues, the agents will handle and resolve the problems according to established procedures.

[0003] To ensure customer service quality and customer experience, most companies conduct quality inspections of their customer service representatives. The typical quality inspection process involves quality inspectors playing back recorded conversations, randomly selecting and scoring them to assess service quality. However, this process is largely manual, resulting in time-consuming playback, potential omissions of inspection items, and low efficiency. Furthermore, with the current surge in inquiries and work orders, the workload for quality inspectors is increasing, necessitating a corresponding increase in the number of personnel, leading to high human resource costs.

[0004] Therefore, there is an urgent need for a work order quality inspection method to reduce manual operations in quality inspection and improve quality inspection efficiency. Summary of the Invention

[0005] In view of this, this application provides a work order quality inspection method, apparatus, equipment and readable storage medium to avoid the problem of spending a lot of time and effort to query and compare information of each carrier one by one.

[0006] To achieve the above objectives, the following solution is proposed:

[0007] A work order quality inspection method, comprising:

[0008] Retrieve conversation information between agents and customers in work orders;

[0009] Identify customer intent based on the session information;

[0010] Determine a combination of quality inspection items that matches the customer's intent, wherein the combination of quality inspection items includes multiple quality inspection items;

[0011] Based on the conversation information and the preset quality inspection item scoring criteria, each quality inspection item in the quality inspection item combination is scored;

[0012] Summarize the scores of each quality inspection item in the aforementioned quality inspection item combination to calculate the total quality inspection score;

[0013] The quality inspection results are determined based on the overall quality inspection score.

[0014] Preferably, the quality inspection item combination includes process quality inspection items and over-authority quality inspection items. The process quality inspection items are used to detect the quality inspection process of the work order, and the over-authority quality inspection items are used to detect whether the operator's operation exceeds the allowed operation range.

[0015] Preferably, the process of scoring the process quality inspection items includes:

[0016] Obtain the quality inspection process from the process quality inspection items, wherein the quality inspection process contains multiple process nodes;

[0017] Based on preset keywords, identify the triggering factors and corresponding quality control factors in the session information;

[0018] Identify the target process node in the quality inspection process that matches the triggering factor;

[0019] Obtain the agent's operation at the target process node;

[0020] Determine whether the agent operation of the target process node matches the target quality inspection factor, wherein the target quality inspection factor is the quality inspection factor corresponding to the trigger factor that matches the target process node;

[0021] If a match is found, the target process node is marked as correct; otherwise, it is marked as incorrect.

[0022] The marking results of all target process nodes in the quality inspection process are summarized, and the process quality inspection item score is calculated.

[0023] Preferably, the process of scoring the over-authority quality inspection items includes:

[0024] Get agent level and all agent operations;

[0025] Determine the operational scope that matches the agent's level;

[0026] Sequentially determine whether each of the agent's operations exceeds the operation range;

[0027] Mark operator actions that exceed the operating range as incorrect, and mark operator actions that do not exceed the operating range as correct;

[0028] By summarizing the marking results of all the agent operations, the score for the over-authority quality inspection item is calculated.

[0029] Preferably, before obtaining the conversation information between the agent and the customer in the work order, the method further includes:

[0030] Determine whether the work order has a quality inspection mark;

[0031] If so, then exit quality inspection;

[0032] If not, proceed with the step of retrieving the conversation information between the agent and the customer from the work order.

[0033] Preferably, after identifying the customer intent based on the session information, the method further includes:

[0034] Determine whether the customer's intent is a preset, quality-inspectable intent;

[0035] If not, then exit the quality inspection.

[0036] If so, proceed with the step of determining the combination of quality inspection items that matches the customer's intent.

[0037] Preferably, the scores of each quality inspection item in the combined quality inspection items are summarized to calculate the total quality inspection score.

[0038] include:

[0039] The scoring results of each quality inspection item are input into a preset quality inspection scoring model to obtain the total quality inspection score output by the quality inspection scoring model.

[0040] The quality inspection scoring model is obtained by using the scores of each sample's quality inspection items as training samples and the total quality inspection score of the samples as sample labels.

[0041] A work order quality inspection device, comprising:

[0042] The acquisition unit is used to acquire the conversation information between the agent and the customer in the work order;

[0043] An intent recognition unit is used to recognize customer intent based on the session information;

[0044] A combination determination unit is used to determine a combination of quality inspection items that matches the customer's intent, wherein a combination of quality inspection items contains multiple quality inspection items;

[0045] The first scoring unit is used to score each quality inspection item in the quality inspection item combination based on the conversation information and the preset quality inspection item scoring criteria.

[0046] The second scoring unit is used to summarize the scoring results of each quality inspection item in the quality inspection item combination and calculate the total quality inspection score.

[0047] The result determination unit is used to determine the quality inspection result based on the total quality inspection score.

[0048] A work order quality inspection device, comprising a memory and a processor;

[0049] The memory is used to store programs;

[0050] The processor is used to execute the program to implement the various steps of the work order quality inspection method described above.

[0051] A readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the various steps of the work order quality inspection method described above.

[0052] As can be seen from the above technical solution, the work order quality inspection method provided in this application obtains the conversation information between the agent and the customer in the work order, identifies the customer's intent based on the conversation information, and determines a combination of quality inspection items that matches the customer's intent. The combination of quality inspection items contains multiple quality inspection items, which are used to inspect various aspects of the work order. The conversation information is inspected according to a preset quality inspection item scoring standard to obtain the score for each quality inspection item in the combination. The scores of each quality inspection item in the combination are summarized to calculate the total quality inspection score, which represents the total score of the work order. Therefore, the quality inspection result of the work order can be determined through the total quality inspection score.

[0053] This application allows for the use of a server to inspect various quality control items in a work order after obtaining the conversation information between the agent and the customer. The server can then determine the quality control result of the work order. During this process, the server can complete the quality control process without the need for manual replay of the conversation, which reduces manual operation and improves quality control efficiency. Attached Figure Description

[0054] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only embodiments of this application. For those skilled in the art, other drawings can be obtained based on the provided drawings without creative effort.

[0055] Figure 1 This is a flowchart of a work order quality inspection method disclosed in an embodiment of this application;

[0056] Figure 2 A flowchart illustrating a return process within a process quality inspection item as exemplified in this application;

[0057] Figure 3 Flowchart of another work order quality inspection method as an example of this application;

[0058] Figure 4 This is a structural block diagram of a work order quality inspection device disclosed in this application;

[0059] Figure 5 This is a hardware structure block diagram of a work order quality inspection device disclosed in an embodiment of this application. Detailed Implementation

[0060] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0061] The following section introduces the application scheme; see [link / reference] Figure 1 , Figure 1 This is a flowchart of a work order quality inspection method disclosed in an embodiment of this application.

[0062] like Figure 1 As shown, the method may include:

[0063] Step S1: Obtain the conversation information between the agent and the customer in the work order.

[0064] Specifically, customers connect to an online agent through the client, creating a work order. They then connect with the agent and communicate, using various methods such as phone calls, voice recordings, or text input. Audio data from phone calls and voice recordings, as well as text input, are captured by the server and stored as part of the current work order's conversation information. This conversation information includes, but is not limited to, voice chat data between the agent and customer, text input in the chat box, and converted text data from voice clips.

[0065] Step S2: Identify customer intent based on the session information.

[0066] Specifically, based on the current interaction information, the system intelligently identifies the customer's intent within the interaction information. Identifying the current intent can be achieved through various methods, such as text matching and intent recognition models. This application does not limit the method used to identify the current customer intent. Customer intent includes scenarios such as returns, exchanges, expedited shipping, and expedited delivery. Different scenarios will correspond to different combinations of quality inspection items. The server can match key information in the dialogue to quickly locate the customer's intent.

[0067] For example, after a customer connects to an agent, they can ask, "Hello, I bought a fan a while ago, can I return it?" The server will recognize the customer's intention as "return" through the voice data.

[0068] Step S3: Determine the combination of quality inspection items that match the customer's intent.

[0069] Specifically, the server supports pre-setting a series of quality inspection item combinations that match customer intent. Each quality inspection item combination contains multiple quality inspection items, such as: service attitude quality inspection items, comprehension ability quality inspection items, service proactivity quality inspection items, reassurance technology quality inspection items, response speed quality inspection items, process quality inspection items, and exceeding authority quality inspection items. Different quality inspection items can be configured according to different customer intent scenarios to form quality inspection item combinations.

[0070] For example, if the customer's intention is "return," a quality inspection item combination can be formed to match the return. This combination includes quality inspection items such as process quality inspection items, over-authorization quality inspection items, and service attitude quality inspection items. Among these, the process quality inspection items are the return process corresponding to the return. Conversely, if the customer's intention is "exchange," a quality inspection item combination can be formed to match the exchange. The process quality inspection items in this combination are the exchange process corresponding to the exchange.

[0071] Step S4: Based on the conversation information and the preset quality inspection item scoring criteria, score each quality inspection item in the quality inspection item combination.

[0072] Specifically, the pre-set scoring criteria for quality inspection items should have different scoring standards for different quality inspection items. For example, for a process quality inspection item where the customer's intention is "return", the full score can be set to 100 points. The conversation information is identified and compared with the return process in the quality inspection process item. If a process node is incorrect, 10 points are deducted, and so on.

[0073] Based on the pre-set quality inspection item scoring criteria, each quality inspection item in the quality inspection item combination is tested to obtain a score for each quality inspection item. The score of each quality inspection item can measure the degree of matching of the corresponding quality inspection item. Generally speaking, the higher the score, the higher the degree of matching of the quality inspection item, the lower the error rate of the quality inspection item, and the better the performance of the work order in the corresponding direction of the quality inspection item.

[0074] Step S5: Summarize the scoring results of each quality inspection item in the combined quality inspection items and calculate the total quality inspection score.

[0075] Specifically, the scores of each quality inspection item in the quality inspection item combination obtained through the above steps are summarized and analyzed and calculated again to obtain the total quality inspection score. The total quality inspection score can measure the overall performance of the work order in multiple aspects. Generally speaking, the higher the score, the higher the degree of matching of the work order in each quality inspection item, the lower the overall error rate of the work order, and the better the overall performance of the work order.

[0076] There are several ways to calculate the total quality inspection score based on the scores of each quality inspection item. For example, it can be calculated by cumulative addition, by weighted average, or by using a trained quality inspection scoring model.

[0077] Step S6: Determine the quality inspection result based on the overall quality inspection score.

[0078] Specifically, the overall quality inspection score can represent the overall performance of the work order in each quality inspection item. Therefore, the quality inspection result can be determined through the overall quality inspection score, which can include quality inspection qualified and quality inspection unqualified.

[0079] There are several ways to determine the quality inspection result based on the overall quality inspection score. For example, a preset quality inspection pass score can be used to determine work orders with an overall quality inspection score higher than the preset pass score as qualified, and other work orders as unqualified. This application does not limit the method of determining the quality inspection result based on the overall quality inspection score.

[0080] This application provides a work order quality inspection method. It acquires conversation information between the agent and the customer within the work order, identifies the customer's intent based on the conversation information, and determines a combination of quality inspection items matching the customer's intent. Each quality inspection item combination contains multiple quality inspection items, which are used to inspect various aspects of the work order. The conversation information is then inspected according to a preset quality inspection item scoring standard to obtain a score for each quality inspection item in the combination. The scores of all quality inspection items in the combination are summarized to calculate the total quality inspection score, which represents the overall score of the work order. Therefore, the quality inspection result of the work order can be determined through the total quality inspection score.

[0081] This application allows for the use of a server to inspect various quality control items in a work order after obtaining the conversation information between the agent and the customer. The server can then determine the quality control result of the work order. During this process, the server can complete the quality control process without the need for manual replay of the conversation, which reduces manual operation and improves quality control efficiency.

[0082] Optionally, in the above embodiments, the process of calculating the total quality inspection score by summarizing the score results of each quality inspection item in step S5 can be achieved by inputting the score results of each quality inspection item into a preset quality inspection scoring model to obtain the total quality inspection score output by the quality inspection scoring model.

[0083] Specifically, the scoring data for each quality inspection item of each sample is collected in advance as training samples, and the total quality inspection score of the sample is obtained as the sample label. The quality inspection scoring model is then trained based on the collected training samples and sample labels.

[0084] After training the quality inspection scoring model, inputting the scores of each quality inspection item into the model will yield the total quality inspection score output by the model.

[0085] In some embodiments of this application, the quality inspection item combination may include process quality inspection items and over-authority quality inspection items.

[0086] Specifically, process quality inspection items can be used to check the quality inspection process of work orders, while over-authority quality inspection items can be used to check whether the operator's operation exceeds the allowed operating range.

[0087] The process quality inspection items include processes that match the customer's intent. Based on the matching process and conversation information, it is determined whether the conversation information matches the steps of the matching process, and whether the agent in the work order provides service in accordance with the steps of the matching process. This is used to inspect the quality inspection process of the work order.

[0088] The over-authority quality check includes the operational scope corresponding to each agent level. Over-authority quality check refers to detecting whether an operation exceeds the allowed authorization range; in this application, it detects whether an agent's operation exceeds the allowed operational scope. Based on the agent's operation and the corresponding operational scope, it is determined whether each operation exceeds the allowed operational scope for that agent level, thereby determining whether the agent's operation exceeds authorization.

[0089] Based on this, combined Figure 2 The process of scoring the process quality inspection items and the process of scoring the over-authority quality inspection items are described in detail. Figure 2 This demonstrates the return process within the quality control items that match the customer's intention to "return" the goods.

[0090] The scoring process for process quality inspection items may include the following seven steps:

[0091] Step 1: Obtain the quality inspection process in the process quality inspection item, which contains multiple process nodes.

[0092] Specifically, based on the customer's intent, a combination of quality inspection items can be obtained that matches the customer's intent. The combination of quality inspection items includes process quality inspection items, and the quality inspection process in the process quality inspection items is also a process that matches the user's intent. The quality inspection process can contain a number of process nodes, and each process node has a corresponding triggering condition.

[0093] like Figure 2 In the return process shown, "Has the shipment been more than 7 days ago?", "Does the 7-day no-reason return policy apply?", and "Is the product amount greater than 500?" are all process nodes in the return process.

[0094] Step 2: Identify the triggering factors and corresponding quality control factors in the session information based on preset keywords.

[0095] Specifically, based on preset keywords, triggering factors and corresponding quality inspection factors in the session information are identified. Triggering factors are keywords that can trigger the detection of the process node, and quality inspection factors are keywords that can be used to judge the process node. If a quality inspection factor matching the triggering factor of the node exists in the session information, the detection of the process node can be triggered; if no quality inspection factor matching the triggering factor of the node exists in the session information, the detection of the process node will not be triggered.

[0096] For example, such as Figure 2 The "Has the shipment exceeded 7 days?" process node shown in the diagram, if the session information is:

[0097] Customer Service Representative: When did you receive your computer?

[0098] Customer: Received it yesterday.

[0099] "When was it received?" will serve as the trigger factor for "Has the shipment been more than 7 days?", triggering the detection of the "Has the shipment been more than 7 days?" process node. The "Received yesterday" message in the conversation information can be used to determine that the shipment has not exceeded 7 days. "Received yesterday" is the quality inspection factor corresponding to this trigger factor.

[0100] Step 3: Identify the target process node in the quality inspection process that matches the triggering factor.

[0101] Specifically, in the above steps, multiple sets of triggering factors and corresponding quality inspection factors can be identified. In the quality inspection process, the target process node that matches the identified triggering factors is queried and matched.

[0102] As in the example in step 2, the trigger factor "when was received" will be used as the trigger factor for the "has been shipped more than 7 days" process in the return process, triggering the "has been shipped more than 7 days" process node. At this time, in the "has been shipped more than 7 days" process node, "when was received" (trigger factor) and "received yesterday" (quality inspection factor) match each other.

[0103] Step 4: Obtain the agent operation at the target process node.

[0104] Step 5: Determine whether the agent operation of the target process node matches the target quality inspection factor, where the target quality inspection factor is the quality inspection factor corresponding to the trigger factor that matches the target process node.

[0105] Specifically, after obtaining the agent's operation for the target process node that matches the trigger factor, it is determined whether the agent's operation matches the target quality inspection factor. The target quality inspection factor is the quality inspection factor corresponding to the trigger factor that matches the target process node. The quality inspection factor is a keyword that can be used to judge the process node; that is, the target process node can be judged through the target quality inspection factor. Based on the judgment result, it is determined whether the agent's operation matches the judgment result.

[0106] For example, in Figure 2 In the return process shown, if the conversation message includes "Your item is less than 500 yuan, I will process your return.", it will trigger the check at the "Is the item amount greater than 500 yuan?" process node. Simultaneously, the quality control factor "less than 500 yuan" will be identified, confirming that the judgment result for this process node is "no," and the return operation should be performed. The agent's actions at the target process node are compared with the actions corresponding to the judgment result. If the actions are the same, it means the agent's actions at the target process node match the target quality control factor; if the actions are different, it means the agent's actions at the target process node do not match the target quality control factor.

[0107] Step 6: If a match is found, mark the target process node as correct; otherwise, mark it as incorrect.

[0108] Specifically, if the agent's operation at the target process node matches the target quality control factor, the target process node is marked as correct; if the agent's operation at the target process node does not match the target quality control factor, the target process node is marked as incorrect. When the target process node is correct, the triggering factor, quality control factor, and agent operation at that target process node are mutually matched.

[0109] For example, in Figure 2 In the return process shown, if the session information is:

[0110] Customer: I bought a computer a while ago, but I don't want it anymore. I'd like to know if I can return it?

[0111] Customer Service Representative: When did you receive your computer?

[0112] Customer: I received it a week ago.

[0113] Customer Service Representative: I'm sorry, but if it's been more than a week, we don't allow returns.

[0114] Meanwhile, the agent registered the work order as non-refundable.

[0115] In the above scenario, at the target process node "Has the shipment exceeded 7 days", the "When was it received" (trigger factor), "Received one week ago" (quality inspection factor), and "No returns allowed" (salesperson operation) match each other, and the target process node is marked as correct.

[0116] Step 7: Summarize the marking results of all target process nodes in the quality inspection process and calculate the process quality inspection item score.

[0117] Specifically, the above methods are used to inspect each process node in the process quality inspection item, obtain the marking results of the obtained process nodes, summarize the marking results of all target process nodes in the quality inspection process, and calculate the process quality inspection item score.

[0118] The process of calculating the process quality inspection item score based on the labeling results of all target process nodes can take several forms. For example, it can be calculated by determining the percentage of correctly labeled target process nodes out of all target process nodes, or by using a trained scoring model. Generally, the closer the number of correctly labeled target process nodes is to the total number of target process nodes, the higher the process quality inspection item score, indicating that the work order's process operation closely matches the pre-set process operation.

[0119] The scoring process for over-authority quality inspection items may include the following five steps:

[0120] Step 1: Obtain agent level and all agent operations.

[0121] Specifically, agents can be categorized into different levels, each with varying permissions and corresponding operational scopes. The system retrieves agent levels and all agent operations for subsequent checks to determine if actions exceed authorized limits.

[0122] Step 2: Determine the operational scope that matches the agent level.

[0123] Specifically, a matching relationship between agent seats and their allowed operating range can be preset. The higher the agent's level, the wider the allowed operating range, allowing them to process higher-level work orders. After obtaining the agent's level, the operating range matching that level can be determined. Newly created work orders can be considered as the lowest level, and agents can upgrade these work orders based on actual circumstances, transferring them to higher-level agents for processing.

[0124] For example, agent seats can be divided into basic and advanced levels, handling basic and advanced work orders respectively. Basic agents can handle work order returns and exchanges under 500 yuan, while advanced agents can handle work order returns and exchanges over 500 yuan. Figure 2In the return process shown, if the value of the goods is greater than 500 yuan, the junior agent can upgrade the work order and transfer it to the senior agent for processing.

[0125] Step 3: Sequentially determine whether each agent operation exceeds the operation range in all the agent operations.

[0126] Specifically, based on the operation range matched to the agent's level, it is sequentially determined whether each agent's operation exceeds the specified operation range. For example, if the product amount in a work order exceeds 500, a junior agent cannot process work order returns or exchanges. If a junior agent performs a return operation on a work order, it exceeds the operation range, and the agent's operation exceeds their authority.

[0127] Step 4: Mark agent operations that exceed the operating range as incorrect, and mark agent operations that do not exceed the operating range as correct.

[0128] Specifically, if an agent's operation exceeds the range of operations corresponding to that agent level, the operation is marked as incorrect; if an agent's operation does not exceed the range of operations corresponding to that agent level, the operation is marked as correct. Each agent operation is checked sequentially, and a corresponding mark is made based on the check results.

[0129] Step 5: Summarize the marking results of all the agent operations and calculate the score for the over-authority quality inspection item.

[0130] Specifically, through the above methods, all agent operations are detected, the marking results of the agent operations are obtained, the marking results of all agent operations are summarized, and the score of the over-authority quality inspection item is calculated.

[0131] The process of calculating the over-authorization quality inspection score based on the marking results of all agent operations can be done in several ways. For example, it can be calculated by determining the percentage of correctly marked agent operations out of all agent operations, or by using a trained scoring model. Generally, the closer the number of correctly marked agent operations is to the total number of agent operations, the higher the over-authorization quality inspection score, indicating fewer instances of over-authorization actions in the work order.

[0132] In some embodiments of this application, considering practical situations, to avoid the server performing duplicate quality checks on the same work order, and also to avoid duplicate quality checks on the same work order by both manual and server inspections when they are performed simultaneously, therefore, as Figure 3 As shown, before step S1, obtaining the conversation information between the agent and the customer in the work order, the following may also be included:

[0133] Step S7: Determine whether the work order has a quality inspection mark.

[0134] Specifically, after a work order is inspected, an inspection result will be obtained. The inspected mark can be the inspection result or a mark added after the inspection.

[0135] The detection process will determine whether a work order has been marked as having undergone quality inspection, and will produce the following two results:

[0136] The first method is to determine if the work order has been marked as having been inspected, which means that the work order has already been inspected manually or by the server and does not need to be inspected again. In this case, the inspection process will be terminated.

[0137] The second approach is to determine if the work order does not have a quality inspection mark, and then proceed to step S1, which involves obtaining the conversation information between the agent and the customer in the work order.

[0138] Optionally, in practical applications, only common customer intent combinations such as returns, exchanges, and expedited shipments are typically pre-set. Therefore, if a work order's customer intent does not have a pre-set corresponding quality inspection item combination, the server cannot perform quality inspection on the work order. Thus, after identifying the customer intent based on the session information in step S2, the process may further include:

[0139] Step S8: Determine whether the customer intent is a preset quality-inspectable intent.

[0140] Specifically, after identifying the customer's intent based on the session information, it is also possible to detect whether the customer's intent is a preset quality-inspectable intent. If the customer's intent is an intent other than the preset quality-inspectable intent, the server cannot perform quality inspection on the work order.

[0141] The detection and judgment of whether the customer's intent is a preset, quality-inspectable intent will produce the following two judgment results:

[0142] The first scenario is that if the judgment result indicates that the customer's intention is not the preset inspectable intention, the server cannot perform quality inspection on the work order and will exit the quality inspection process.

[0143] The second approach is to determine the customer's intent as a pre-defined quality-inspectable intent if the customer's intent is the same as the pre-defined quality-inspectable intent. If the customer's intent is the same as the pre-defined quality-inspection intent, then proceed to step S3, which involves determining the combination of quality inspection items that match the customer's intent.

[0144] The work order quality inspection device provided in the embodiments of this application is described below. The work order quality inspection device described below can be referred to in correspondence with the work order quality inspection method described above.

[0145] See Figure 4 , Figure 4 This is a schematic diagram of the structure of a work order quality inspection device disclosed in an embodiment of this application.

[0146] like Figure 4 As shown, the device may include:

[0147] Acquisition unit 110 is used to acquire conversation information between the agent and the customer in the work order;

[0148] The intent recognition unit 120 is used to recognize the customer's intent based on the session information;

[0149] The combination determination unit 130 is used to determine a combination of quality inspection items that matches the customer's intention, wherein a combination of quality inspection items contains multiple quality inspection items;

[0150] The first scoring unit 140 is used to score each quality inspection item in the quality inspection item combination according to the conversation information and the preset quality inspection item scoring criteria.

[0151] The second scoring unit 150 is used to summarize the scoring results of each quality inspection item in the quality inspection item combination and calculate the total quality inspection score.

[0152] Result determination unit 160 is used to determine the quality inspection result based on the total quality inspection score.

[0153] As can be seen from the above technical solution, the work order quality inspection device provided in this application obtains the conversation information between the agent and the customer in the work order, identifies the customer's intent based on the conversation information, and determines a combination of quality inspection items that matches the customer's intent. The combination of quality inspection items contains multiple quality inspection items, which are used to inspect various aspects of the work order. The conversation information is inspected according to a preset quality inspection item scoring standard to obtain a score for each quality inspection item in the combination. The scores of each quality inspection item in the combination are summarized to calculate the total quality inspection score, which represents the total score of the work order. Therefore, the quality inspection result of the work order can be determined through the total quality inspection score.

[0154] This application allows for the use of a server to inspect various quality control items in a work order after obtaining the conversation information between the agent and the customer. The server can then determine the quality control result of the work order. During this process, the server can complete the quality control process without the need for manual replay of the conversation, which reduces manual operation and improves quality control efficiency.

[0155] Optionally, the above-mentioned quality inspection item combination includes process quality inspection items and over-authority quality inspection items. The process quality inspection items are used to detect the quality inspection process of the work order, and the over-authority quality inspection items are used to detect whether the agent's operation exceeds the allowed operation range.

[0156] Optionally, the first scoring unit mentioned above may include a process quality inspection subunit;

[0157] The process by which the above-mentioned process quality inspection subunit scores the process quality inspection items may include:

[0158] Obtain the quality inspection process from the process quality inspection items, wherein the quality inspection process contains multiple process nodes;

[0159] Based on preset keywords, identify the triggering factors and corresponding quality control factors in the session information;

[0160] Identify the target process node in the quality inspection process that matches the triggering factor;

[0161] Obtain the agent's operation at the target process node;

[0162] Determine whether the agent operation of the target process node matches the target quality inspection factor, wherein the target quality inspection factor is the quality inspection factor corresponding to the trigger factor that matches the target process node;

[0163] If a match is found, the target process node is marked as correct; otherwise, it is marked as incorrect.

[0164] The marking results of all target process nodes in the quality inspection process are summarized, and the process quality inspection item score is calculated.

[0165] Optionally, the first scoring unit mentioned above may include an over-authority quality inspection subunit;

[0166] The process by which the above-mentioned over-authority quality inspection subunit scores the process quality inspection items may include:

[0167] Get agent level and all agent operations;

[0168] Determine the operational scope that matches the agent's level;

[0169] Sequentially determine whether each of the agent's operations exceeds the operation range;

[0170] Mark operator actions that exceed the operating range as incorrect, and mark operator actions that do not exceed the operating range as correct;

[0171] By summarizing the marking results of all the agent operations, the score for the over-authority quality inspection item is calculated.

[0172] Optionally, the work order quality inspection device of this application may further include:

[0173] The marking detection unit is used to determine whether the work order has been marked as having been inspected. If so, the quality inspection is terminated.

[0174] The aforementioned acquisition unit can also be used to acquire the conversation information between the agent and the customer in a work order when the work order does not have a quality inspection mark.

[0175] Optionally, the work order quality inspection device of this application may further include:

[0176] The intent detection unit is used to determine whether the customer intent is a preset quality-inspectable intent; if not, the quality inspection is terminated.

[0177] The aforementioned acquisition unit can also be used to determine a combination of quality inspection items that matches the customer's intent when the customer's intent is not a preset quality-inspectable intent.

[0178] The work order quality inspection device provided in this application embodiment can be applied to work order quality inspection equipment, such as PC terminals, cloud platforms, servers, and server clusters. Optionally, Figure 5 The hardware structure block diagram of the work order quality inspection equipment is shown below. Figure 5 The hardware structure of the work order quality inspection equipment may include: at least one processor 1, at least one communication interface 2, at least one memory 3, and at least one communication bus 4.

[0179] In this embodiment of the application, the number of processor 1, communication interface 2, memory 3, and communication bus 4 is at least one, and processor 1, communication interface 2, and memory 3 communicate with each other through communication bus 4;

[0180] Processor 1 may be a central processing unit (CPU), an application-specific integrated circuit (ASIC), or one or more integrated circuits configured to implement embodiments of the present invention.

[0181] Memory 3 may include high-speed RAM, and may also include non-volatile memory, such as at least one disk storage device;

[0182] The memory stores a program, which the processor can call. The program is used for:

[0183] Retrieve conversation information between agents and customers in work orders;

[0184] Identify customer intent based on the session information;

[0185] Determine a combination of quality inspection items that matches the customer's intent, wherein the combination of quality inspection items includes multiple quality inspection items;

[0186] Based on the conversation information and the preset quality inspection item scoring criteria, each quality inspection item in the quality inspection item combination is scored;

[0187] Summarize the scores of each quality inspection item in the aforementioned quality inspection item combination to calculate the total quality inspection score;

[0188] The quality inspection results are determined based on the overall quality inspection score.

[0189] Optionally, the refined and extended functions of the program can be referred to the above description.

[0190] This application embodiment also provides a readable storage medium that can store a program suitable for execution by a processor, the program being used for:

[0191] Retrieve conversation information between agents and customers in work orders;

[0192] Identify customer intent based on the session information;

[0193] Determine a combination of quality inspection items that matches the customer's intent, wherein the combination of quality inspection items includes multiple quality inspection items;

[0194] Based on the conversation information and the preset quality inspection item scoring criteria, each quality inspection item in the quality inspection item combination is scored;

[0195] Summarize the scores of each quality inspection item in the aforementioned quality inspection item combination to calculate the total quality inspection score;

[0196] The quality inspection results are determined based on the overall quality inspection score.

[0197] Optionally, the refined and extended functions of the program can be referred to the above description.

[0198] Finally, it should be noted that in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0199] The various embodiments in this specification are described in a progressive manner, with each embodiment focusing on the differences from other embodiments. The same or similar parts between the various embodiments can be referred to each other.

[0200] The above description of the disclosed embodiments enables those skilled in the art to make or use this application. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of this application. Therefore, this application is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A work order quality inspection method, characterized in that, The method comprises the following steps: obtaining session information of a session between an agent and a customer in a work order; identifying a customer intention according to the session information; determining a quality inspection item combination matched with the customer intention, wherein the quality inspection item combination comprises a process quality inspection item and a super permission quality inspection item, the process quality inspection item is used for detecting a quality inspection process of the work order, and the super permission quality inspection item is used for detecting whether an operation of the agent exceeds an allowed operation range; scoring each quality inspection item in the quality inspection item combination according to the session information and a preset quality inspection item scoring standard; summarizing scoring results of each quality inspection item in the quality inspection item combination to obtain a quality inspection total score; determining a quality inspection result according to the quality inspection total score; the process of scoring the process quality inspection item comprises the following steps: obtaining a quality inspection process in the process quality inspection item, wherein the quality inspection process comprises a plurality of process nodes; identifying a trigger factor and a corresponding quality inspection factor in the session information according to a preset keyword; determining a target process node matched with the trigger factor in the quality inspection process; obtaining an agent operation at the target process node; judging whether the agent operation at the target process node matches a target quality inspection factor, wherein the target quality inspection factor is a quality inspection factor corresponding to the target process node; if the agent operation matches the target quality inspection factor, marking the target process node as correct, otherwise marking the target process node as incorrect; summarizing marking results of all target process nodes in the quality inspection process to obtain a process quality inspection item score; the process of scoring the super permission quality inspection item comprises the following steps: obtaining an agent level and all agent operations; determining an operation range matched with the agent level; judging whether each agent operation in the all agent operations exceeds the operation range; marking an agent operation exceeding the operation range as incorrect, and marking an agent operation not exceeding the operation range as correct; summarizing marking results of the all agent operations to obtain a super permission quality inspection item score.

2. The method of claim 1, wherein, Before the step of obtaining the session information of the session between the agent and the customer in the work order, the method further comprises the following steps: judging whether the work order has a quality inspection marked; if yes, exiting the quality inspection; if no, performing the step of obtaining the session information of the session between the agent and the customer in the work order.

3. The method of claim 1, wherein, After the step of identifying the customer intention according to the session information, the method further comprises the following steps: judging whether the customer intention is a preset quality inspection intention; if no, exiting the quality inspection; if yes, performing the step of determining the quality inspection item combination matched with the customer intention.

4. The method of claim 1, wherein, The step of summarizing the scoring results of each quality inspection item in the quality inspection item combination to obtain the quality inspection total score comprises the following steps: inputting the scoring results of each quality inspection item into a preset quality inspection scoring model to obtain the quality inspection total score output by the quality inspection scoring model; the quality inspection scoring model is obtained by pre-training with each sample quality inspection item score as a training sample and with a sample quality inspection total score as a sample label.

5. A work order quality inspection device, comprising: The method comprises the following steps: an obtaining unit, configured to obtain session information of a session between an agent and a customer in a work order; an intention identifying unit, configured to identify a customer intention according to the session information; The combination determining unit is configured to determine a combination of quality inspection items matched with the customer intention, wherein the combination of quality inspection items comprises a process quality inspection item and a super permission quality inspection item, the process quality inspection item is configured to detect a quality inspection process of the work order, and the super permission quality inspection item is configured to detect whether the operation of the agent exceeds an allowed operation range; The first scoring unit is configured to score each quality inspection item of the combination of quality inspection items according to the conversation information and a preset quality inspection item scoring standard; The second scoring unit is configured to aggregate the scoring results of each quality inspection item of the combination of quality inspection items to obtain a total quality inspection score; The result determining unit is configured to determine a quality inspection result according to the total quality inspection score; The process of scoring the process quality inspection item comprises: Obtaining a quality inspection process in the process quality inspection item, wherein the quality inspection process comprises a plurality of process nodes; Identifying a trigger factor and a corresponding quality inspection factor in the conversation information according to a preset keyword; Determining a target process node matched with the trigger factor in the quality inspection process; Obtaining an agent operation at the target process node; Judging whether the agent operation at the target process node matches a target quality inspection factor, wherein the target quality inspection factor is a quality inspection factor corresponding to the target process node; If the target process node is correct, marking the target process node as correct, otherwise marking the target process node as incorrect; Aggregating the marking results of all target process nodes in the quality inspection process to obtain a process quality inspection item score; The process of scoring the super permission quality inspection item comprises: Obtaining an agent level and all agent operations; Determining an operation range matched with the agent level; Judging whether each agent operation in the all agent operations exceeds the operation range; Marking the agent operation exceeding the operation range as incorrect, and marking the agent operation not exceeding the operation range as correct; Aggregating the marking results of the all agent operations to obtain a super permission quality inspection item score.

6. A work order quality inspection apparatus, characterized by, The computer program is executed by the processor to implement the steps of the work order quality inspection method. The computer program is executed by the processor to implement the steps of the work order quality inspection method. ​ 7. A readable storage medium, having stored thereon a computer program, characterized in that, ​

Images