Customer service dialogue data processing method, device, equipment, medium and program product

Abstract

This application provides a customer service dialogue data processing method applicable to the blockchain field. The method includes: responding to a data processing instruction, determining target customer service dialogue data, and generating real-time verification information based on the target customer service dialogue data; acquiring first verification information pre-stored in the target blockchain, and determining a first verification result based on the first verification information and the real-time verification information, wherein the first verification information is generated and stored in the target blockchain after the initial collection of the target customer service dialogue data; responding to the first verification result indicating that the target customer service dialogue data has not been modified, processing the target customer service dialogue data using a large language model to obtain a quality inspection result; generating second verification information based on the quality inspection result, and uploading the second verification information to the target blockchain for evidence storage. This application also provides a customer service dialogue data processing device, equipment, medium, and program product.

Description

Technical Field

[0001] This application relates to the field of blockchain, specifically to a customer service dialogue data processing method, apparatus, device, medium, and program product. Background Technology

[0002] In the customer service system of financial services, customer service dialogues are a crucial record of the service process and a reflection of service quality. To ensure service compliance and improve service levels, it is typically necessary to conduct quality checks and analyses on customer service dialogue data. In some cases, this relies on post-event manual sampling checks or automated screening based on pre-set rule engines. However, manual sampling has limited coverage and is inefficient, while rule engines are rigid in handling the diversity and complexity of natural language, making it difficult to effectively identify semantic violations. Furthermore, the processing and results of these methods lack effective anti-tampering mechanisms, making it difficult to effectively verify the authenticity and completeness of the quality inspection results. In the event of disputes or when audits are required, it is impossible to provide convincing, traceable, and objective evidence.

[0003] Therefore, the lack of reliable guarantees for data integrity will affect the credibility of subsequent intelligent analysis using automated models. Related technologies for processing customer service dialogue data suffer from inefficiency, unreliable results, and difficulty in verification. Summary of the Invention

[0004] In view of the above problems, this application provides a customer service dialogue data processing method, apparatus, equipment, medium and program product to improve the reliability of customer service dialogue data processing.

[0005] According to a first aspect of this application, a customer service dialogue data processing method is provided, comprising: responding to a data processing instruction, determining target customer service dialogue data, and generating real-time verification information based on the target customer service dialogue data, wherein the data processing instruction is used to trigger processing of the specified customer service dialogue data; acquiring first verification information pre-stored in a target blockchain, and determining a first verification result based on the first verification information and the real-time verification information, wherein the first verification information is verification information generated and stored in the target blockchain after the target customer service dialogue data has been initially collected; responding to the first verification result indicating that the target customer service dialogue data has not been modified, processing the target customer service dialogue data using a large language model to obtain a quality inspection result; generating second verification information based on the quality inspection result, and uploading the second verification information to the target blockchain for evidence storage.

[0006] According to an embodiment of this application, a large language model is used to process target customer service dialogue data to obtain quality inspection results, including: performing compliance quality inspection on the target customer service dialogue data to obtain a result for identifying illegal expressions, wherein the result for identifying illegal expressions indicates whether the dialogue content corresponding to the dialogue data complies with predefined regulatory rules; performing service quality assessment on the target customer service dialogue data to obtain a service quality score, wherein the service quality score; wherein the quality inspection results include the result for identifying illegal expressions and the service quality score.

[0007] According to an embodiment of this application, the process of processing target customer service dialogue data using a large language model to obtain quality inspection results further includes: determining target dialogue data from the target customer service dialogue data based on the results of identifying illegal expressions and service quality scores, wherein the target dialogue data is a content segment in the target customer service dialogue data that does not contain illegal expressions and whose corresponding service quality score is higher than a preset threshold; and generating knowledge suggestion entries based on the target dialogue data, wherein the knowledge suggestion entries include question descriptions and answer descriptions.

[0008] According to an embodiment of this application, after processing the target customer service dialogue data using a large language model to obtain a quality inspection result in response to the first verification result indicating that the target customer service dialogue data has not been modified, the method further includes: generating third verification information based on knowledge suggestion entries; and uploading the third verification information to the target blockchain for storage.

[0009] According to an embodiment of this application, before determining the target customer service dialogue data in response to a data processing instruction and generating real-time verification information based on the target customer service dialogue data, the method further includes: in response to acquiring the target customer service dialogue data at the initial acquisition time, converting the voice data in the target customer service dialogue data into initial text data; performing desensitization processing on the initial text data to obtain target text data; generating first verification information based on the target text data, and uploading the first verification information to the target blockchain for storage.

[0010] According to an embodiment of this application, generating second verification information based on the quality inspection processing result includes: generating a target hash value based on the quality inspection processing result; obtaining target metadata corresponding to the quality inspection processing result, wherein the target metadata includes at least a processing timestamp; and generating second verification information based on the target hash value and the target metadata.

[0011] According to an embodiment of this application, the method further includes: obtaining a target verification request, wherein the target verification request includes off-chain data to be verified; generating target verification information based on the off-chain data to be verified; obtaining on-chain verification information corresponding to the off-chain data to be verified, wherein the on-chain verification information includes first verification information or second verification information; and determining a target verification result based on the target verification information and the on-chain verification information.

[0012] The second aspect of this application provides a customer service dialogue data processing device, comprising: a first processing module, configured to, in response to a data processing instruction, determine target customer service dialogue data and generate real-time verification information based on the target customer service dialogue data, wherein the data processing instruction is used to trigger processing of the specified customer service dialogue data; a second processing module, configured to acquire first verification information pre-stored in a target blockchain and determine a first verification result based on the first verification information and the real-time verification information, wherein the first verification information is verification information generated and stored in the target blockchain after the target customer service dialogue data has been initially collected; a third processing module, configured to, in response to the first verification result indicating that the target customer service dialogue data has not been modified, process the target customer service dialogue data using a large language model to obtain a quality inspection result; and a fourth processing module, configured to, based on the quality inspection result, generate second verification information and upload the second verification information to the target blockchain for evidence storage.

[0013] A third aspect of this application provides an electronic device comprising: one or more processors; and a memory for storing one or more computer programs, wherein the one or more processors execute the one or more computer programs to implement the steps of the method described above.

[0014] A fourth aspect of this application also provides a computer-readable storage medium having a computer program or instructions stored thereon, which, when executed by a processor, implement the steps of the above-described method.

[0015] The fifth aspect of this application also provides a computer program product, including a computer program or instructions that, when executed by a processor, implement the steps of the above-described method. Attached Figure Description

[0016] The above-mentioned contents, other objects, features and advantages of this application will become clearer from the following description of embodiments with reference to the accompanying drawings, in which:

[0017] Figure 1 The illustration shows an application scenario diagram of the customer service dialogue data processing method, apparatus, device, medium, and program product according to embodiments of this application;

[0018] Figure 2 A flowchart illustrating a customer service dialogue data processing method according to an embodiment of this application is shown schematically.

[0019] Figure 3 The schematic diagram illustrates the principle of a customer service dialogue data processing method according to an embodiment of this application;

[0020] Figure 4 This schematically illustrates a structural block diagram of a customer service dialogue data processing apparatus according to an embodiment of this application; and

[0021] Figure 5 A block diagram schematically illustrates an electronic device suitable for implementing a customer service dialogue data processing method according to an embodiment of this application. Detailed Implementation

[0022] The embodiments of this application will now be described with reference to the accompanying drawings. However, it should be understood that these descriptions are exemplary only and are not intended to limit the scope of this application. In the following detailed description, numerous specific details are set forth to provide a thorough understanding of the embodiments of this application for ease of explanation. However, it will be apparent that one or more embodiments may be implemented without these specific details. Furthermore, descriptions of well-known structures and technologies are omitted in the following description to avoid unnecessarily obscuring the concepts of this application.

[0023] The terminology used herein is for the purpose of describing particular embodiments only and is not intended to limit the scope of this application. The terms “comprising,” “including,” etc., as used herein indicate the presence of features, steps, operations, and / or components, but do not exclude the presence or addition of one or more other features, steps, operations, or components.

[0024] All terms used herein (including technical and scientific terms) have the meanings commonly understood by those skilled in the art, unless otherwise defined. It should be noted that the terms used herein are to be interpreted in a manner consistent with the context of this specification, and not in an idealized or overly rigid way.

[0025] When using expressions such as "at least one of A, B and C", they should generally be interpreted in accordance with the meaning that is commonly understood by those skilled in the art (e.g., "a system having at least one of A, B and C" should include, but is not limited to, a system having A alone, a system having B alone, a system having C alone, a system having A and B, a system having A and C, a system having B and C, and / or a system having A, B and C, etc.).

[0026] In the technical solution of this application, the user information (including but not limited to user personal information, user image information, user device information, such as location information) and data (including but not limited to data used for analysis, stored data, and displayed data) involved are all information and data authorized by the user or fully authorized by all parties. Furthermore, the collection, storage, use, processing, transmission, provision, disclosure, and application of related data all comply with relevant laws, regulations, and standards, take necessary confidentiality measures, do not violate public order and good morals, and provide corresponding operation entry points for users to choose to authorize or refuse.

[0027] In scenarios involving automated decision-making using personal information, the methods, devices, and systems provided in this application all offer users corresponding entry points for choosing to agree to or reject the automated decision-making results. If the user chooses to reject, the process proceeds to the expert decision-making stage. Here, "automated decision-making" refers to the activity of automatically analyzing and evaluating an individual's behavioral habits, interests, or economic, health, and credit status through computer programs, and then making a decision. Here, "expert decision-making" refers to the activity of making decisions by personnel who specialize in a particular field, possess specialized experience, knowledge, and skills, and have reached a certain level of professional expertise.

[0028] Figure 1 The illustration shows an application scenario diagram of the customer service dialogue data processing method, apparatus, device, medium, and program product according to embodiments of this application.

[0029] like Figure 1 As shown, application scenario 100 according to this embodiment may include the financial technology field. Network 104 is used as a medium to provide a communication link between the first terminal device 101, the second terminal device 102, the third terminal device 103, and the server 105. Network 104 may include various connection types, such as wired or wireless communication links or fiber optic cables, etc.

[0030] Users can use the first terminal device 101, the second terminal device 102, and the third terminal device 103 to interact with the server 105 via the network 104 to receive or send messages, etc. Various communication client applications can be installed on the first terminal device 101, the second terminal device 102, and the third terminal device 103, such as shopping applications, web browser applications, search applications, instant messaging tools, email clients, social media platform software, etc. (for example only).

[0031] The first terminal device 101, the second terminal device 102, and the third terminal device 103 can be various electronic devices with displays and support web browsing, including but not limited to smartphones, tablets, laptops, and desktop computers.

[0032] Server 105 can be a server that provides various services, such as a backend management server that supports websites browsed by users using the first terminal device 101, the second terminal device 102, and the third terminal device 103 (this is just an example). The backend management server can analyze and process data such as received user requests, and feed back the processing results (such as web pages, information, or data obtained or generated according to user requests) to the terminal devices.

[0033] It should be noted that the customer service dialogue data processing method provided in this application embodiment can generally be executed by server 105. Correspondingly, the customer service dialogue data processing device provided in this application embodiment can generally be located in server 105. The customer service dialogue data processing method provided in this application embodiment can also be executed by a server or server cluster that is different from server 105 and capable of communicating with the first terminal device 101, the second terminal device 102, the third terminal device 103, and / or server 105. Correspondingly, the customer service dialogue data processing device provided in this application embodiment can also be located in a server or server cluster that is different from server 105 and capable of communicating with the first terminal device 101, the second terminal device 102, the third terminal device 103, and / or server 105.

[0034] It should be understood that Figure 1 The number of terminal devices, networks, and servers shown is merely illustrative. Depending on implementation needs, any number of terminal devices, networks, and servers can be included.

[0035] The following will be based on Figure 1 The described scene, through Figures 2-5 The customer service dialogue data processing method according to the embodiments of this application will be described in detail.

[0036] Figure 2 A flowchart illustrating a customer service dialogue data processing method according to an embodiment of this application is shown.

[0037] like Figure 2 As shown, the customer service dialogue data processing method of this embodiment includes operations S210 to S240, and the customer service dialogue data processing method can be executed by the server.

[0038] In operation S210, in response to the data processing instruction, the target customer service dialogue data is determined, and real-time verification information is generated based on the target customer service dialogue data. The data processing instruction is used to trigger the processing of the specified customer service dialogue data.

[0039] In the embodiments of this application, a data processing instruction refers to a command that triggers the system to begin quality inspection analysis of a specific customer service conversation. Data processing instructions can be automatically generated by a system scheduled task or manually triggered by an administrator. For example, an instruction triggered by a system scheduled task, or an analysis instruction manually initiated by an administrator for a suspicious conversation.

[0040] In the embodiments of this application, target customer service dialogue data refers to the customer service dialogue record to be processed as specified by the data processing instruction. Target customer service dialogue data can be stored in the system database in audio or text format. For example, target customer service dialogue data can be a complete audio or text recording of a conversation between customer service representative number 888 and a customer at 10:00 AM on May 27, 2024.

[0041] In the embodiments of this application, real-time verification information refers to cryptographic information generated in real time based on the target customer service dialogue data acquired at the current moment, used to verify data integrity. Real-time verification information can be a hash value obtained by performing a hash operation on the target customer service dialogue data.

[0042] In the embodiments of this application, in response to a data processing instruction, target customer service dialogue data is determined, and real-time verification information is generated based on the target customer service dialogue data. After receiving the data processing instruction, the system first locates the target customer service dialogue data to be analyzed and immediately calculates its current real-time verification information. The purpose of the steps in this embodiment is to capture the immediate state of the data before analysis.

[0043] For example, the system receives a command to analyze "the conversation with customer service session ID 123". The system then locates the text record of that conversation and uses a hash algorithm to calculate its hash value.

[0044] In operation S220, the first verification information pre-stored in the target blockchain is obtained, and the first verification result is determined based on the first verification information and the real-time verification information. The first verification information is the verification information generated and stored in the target blockchain after the initial collection of the target customer service dialogue data is completed.

[0045] In the embodiments of this application, the target blockchain refers to a specific blockchain network used for storing evidence data. For example, the target blockchain can be a consortium blockchain jointly maintained by banks, regulatory agencies, etc.

[0046] In the embodiments of this application, the first verification information refers to the data digest that has been generated and stored in the target blockchain when the target customer service dialogue data is initially generated and collected by the system. The first verification information represents the original state of the data at the time of acquisition.

[0047] In the embodiments of this application, the first verification result refers to the conclusion drawn by comparing whether the real-time verification information is consistent with the first verification information, which is used to determine whether the target customer service dialogue data has been tampered with from the time the data is first stored to the current processing time.

[0048] In the embodiments of this application, first verification information pre-stored in the target blockchain is obtained, and a first verification result is determined based on the first verification information and real-time verification information. The system retrieves the previously stored first verification information of the dialogue data from the blockchain and compares it with the newly generated real-time verification information. If the two are completely consistent, it proves that the data has not been modified since its storage and is trustworthy; otherwise, it indicates that the data has been corrupted.

[0049] For example, the system retrieves the first verification information of session 123 from the blockchain, which matches the newly calculated real-time hash. The first verification result is "data has not been modified".

[0050] In operation S230, in response to the first verification result indicating that the target customer service dialogue data has not been modified, the target customer service dialogue data is processed using a large language model to obtain the quality inspection processing result.

[0051] In the embodiments of this application, the large language model refers to an artificial intelligence model with powerful natural language understanding and generation capabilities. The quality inspection result refers to the structured conclusions output by the large language model after analyzing the target customer service dialogue data. For example, the quality inspection result may include compliance status, service quality score, etc.

[0052] In the embodiments of this application, in response to the first verification result indicating that the target customer service dialogue data has not been modified, the target customer service dialogue data is processed using a large language model to obtain a quality inspection result. Only if the data integrity passes verification will the large language model be initiated for in-depth analysis. The method of this embodiment ensures that the intelligent model's analysis is based on real, uncontaminated data, thereby guaranteeing the reliability of the analysis results.

[0053] For example, after confirming the reliability of the data in session 123, the system inputs it into a large language model. The model then analyzes the data and outputs the quality control results.

[0054] In operation S240, based on the quality inspection results, a second verification message is generated and uploaded to the target blockchain for storage.

[0055] In the embodiments of this application, the second verification information refers to the data summary generated based on the quality inspection results, which is used to uniquely identify and verify the analysis results.

[0056] In the embodiments of this application, second verification information is generated based on the quality inspection results, and this second verification information is uploaded to the target blockchain for notarization. The steps in this embodiment are the result notarization stage. Second verification information is generated based on the analysis results produced by the large language model and permanently recorded on the blockchain. This makes the analysis results themselves immutable and traceable, providing authoritative evidence for subsequent queries, audits, or dispute resolution.

[0057] For example, the system performs a hash operation on the quality inspection results to obtain second verification information, packages the second verification information into a transaction, and writes it into the target blockchain.

[0058] Through the embodiments of this application, by first storing the verification information (first verification information) of the original data on the blockchain, and performing data integrity verification before processing, the authenticity of the underlying data for subsequent intelligent analysis is ensured. Then, the verification information (second verification information) of the quality inspection results generated by the large language model is stored on the blockchain, making the results verifiable and non-repudiable. The method of this embodiment constructs a trusted verification chain throughout the entire process from the data source to the processing result, effectively solving the key problems in traditional customer service quality inspection, such as data susceptibility to tampering, low credibility of analysis results, and lack of authoritative audit evidence, thereby improving the fairness and reliability of customer service dialogue data processing methods.

[0059] In some embodiments, a large language model is used to process target customer service dialogue data to obtain quality inspection results, including: performing compliance quality inspection on the target customer service dialogue data to obtain non-compliant expression identification results, wherein the non-compliant expression identification results characterize whether the dialogue content corresponding to the dialogue data conforms to predefined regulatory rules; performing service quality assessment on the target customer service dialogue data to obtain a service quality score, wherein the service quality score; wherein the quality inspection results include non-compliant expression identification results and service quality scores.

[0060] In the embodiments of this application, compliance inspection refers to the process of checking the content of customer service conversations for compliance in accordance with relevant laws and regulations of the financial industry and internal regulatory rules. The result of identifying non-compliant expressions refers to the conclusion drawn from analyzing the specific language content and its type that violates preset rules.

[0061] In the embodiments of this application, compliance quality checks are performed on the target customer service dialogue data to obtain the results of identifying inappropriate statements. The steps of this embodiment require the large language model to scan the target customer service dialogue data based on its learned regulatory knowledge to identify whether there are inappropriate statements such as illegal promises or misleading sales.

[0062] For example, the model analyzes the dialogue text and if it detects unauthorized promises such as "the principal is absolutely safe" or "the returns will definitely double," it generates a violation statement identification result, such as: "Violation statement detected: promise to guarantee principal and returns."

[0063] In the embodiments of this application, service quality assessment refers to the evaluation process of the overall service level, including the service attitude and problem-solving efficiency of customer service personnel. Service quality score refers to the numerical value output after quantifying the aforementioned service level.

[0064] In the embodiments of this application, service quality assessment is performed on the target customer service dialogue data to obtain a service quality score. Service quality assessment refers to the process of analyzing and quantifying the service level of customer service personnel from multiple dimensions using a large language model. The service quality score is the final quantitative output of the service quality assessment process, usually a single score, used to comprehensively reflect the service level of this customer service dialogue.

[0065] In the embodiments of this application, the large language model can extract features from target customer service dialogue data, such as problem-solving efficiency, answer accuracy, information completeness, and polite language usage, and perform a comprehensive evaluation based on predetermined standards to output a quantitative score.

[0066] For example, in a complaint handling dialogue, the large language model assessment found that the customer service representative provided accurate answers, clear steps, proactively offered solutions, and used polite language. Based on a 0-100 point scoring system, the model assigned a service quality score of 92 points to this dialogue.

[0067] Through the embodiments of this application, by limiting the parallel execution of compliance quality inspection and good service assessment using a large language model, multi-dimensional and integrated intelligent analysis of customer service dialogues is achieved, significantly improving the automation level and comprehensive analysis capabilities of the quality inspection process, and overcoming the limitations of traditional single-function tools.

[0068] In some embodiments, the process of using a large language model to process target customer service dialogue data to obtain quality inspection results further includes: determining target dialogue data from the target customer service dialogue data based on the results of identifying illegal expressions and service quality scores, wherein the target dialogue data is a content segment in the target customer service dialogue data that does not contain illegal expressions and whose corresponding service quality score is higher than a preset threshold; and generating knowledge suggestion entries based on the target dialogue data, wherein the knowledge suggestion entries include question descriptions and answer descriptions.

[0069] In the embodiments of this application, target dialogue data refers to content fragments that meet specific quality requirements and are selected from the target customer service dialogue data to be analyzed. The preset threshold is a minimum passing score set for service quality rating, used to quantitatively measure whether the service quality meets the "excellent" standard.

[0070] In the embodiments of this application, target dialogue data is determined from target customer service dialogue data based on the results of violation statement identification and service quality scores. This embodiment's steps constitute a quality filtering stage in knowledge mining. The system does not blindly extract knowledge from all dialogues, but rather filters based on two key indicators: violation statement identification results and service quality scores. Any dialogue content containing violation statements must be excluded, and, while adhering to compliance, only dialogue segments with service quality scores higher than a preset threshold (e.g., scores above 90) are selected as target dialogue data. The steps of this embodiment ensure that the subsequently generated knowledge entries originate from standardized and efficient service practices.

[0071] For example, suppose there is a segment of customer service dialogue data in the target database where the violation statement is identified as "no violation" and the service quality score is 95. Since 95 is higher than the preset threshold of 90 and there is no violation, this segment of dialogue data is identified as the target dialogue data.

[0072] In the embodiments of this application, a knowledge suggestion entry refers to a structured data object used to represent a complete question-answer pair and is candidate content for knowledge base updates. A question description refers to a standardized summary of user questions extracted from the target dialogue data. An answer description refers to a standardized summary of customer service answers extracted from the target dialogue data.

[0073] In the embodiments of this application, knowledge suggestion entries are generated based on target dialogue data. A large language model understands and analyzes the selected high-quality target dialogue data, accurately extracting the user's core questions and summarizing the effective solutions provided by customer service, resulting in a knowledge suggestion entry containing a standard question description and a standard answer description.

[0074] Through the embodiments of this application, by setting clear filtering criteria, compliant and high-quality content is automatically selected from massive dialogues as knowledge sources, and further structured into standard question-and-answer pairs. The method of this embodiment automates and controls the quality of the knowledge mining process, ensuring the accuracy and practicality of the generated knowledge suggestions.

[0075] In some embodiments, after processing the target customer service dialogue data using a large language model to obtain a quality inspection result in response to the first verification result indicating that the target customer service dialogue data has not been modified, the method further includes: generating third verification information based on knowledge suggestion entries; and uploading the third verification information to the target blockchain for storage.

[0076] In the embodiments of this application, the third verification information refers to a data digest specifically generated for the content of a knowledge suggestion entry. The third verification information is a unique and tamper-proof digital fingerprint of the knowledge suggestion entry content.

[0077] In the embodiments of this application, third verification information is generated based on knowledge suggestion entries. The steps of this embodiment aim to create a unique identifier for each generated knowledge suggestion entry. The system performs a hash operation on the knowledge suggestion entry (containing standardized question and answer descriptions) to generate a unique hash value, i.e., the third verification information. This operation can transform knowledge content into a form that can be cryptographically verified.

[0078] In the embodiments of this application, evidence preservation refers to writing third-party verification information as a transaction into the blockchain, utilizing its immutable characteristics to achieve permanent and traceable records.

[0079] In the embodiments of this application, third verification information is uploaded to the target blockchain for notarization. The steps of this embodiment permanently and irrefutably anchor the "digital fingerprint" of the knowledge suggestion entry to the trusted target blockchain. By uploading the third verification information to the blockchain, the existence, content integrity, and generation time of the knowledge suggestion are all proven by the target blockchain.

[0080] Through the embodiments of this application, by generating independent on-chain evidence (third verification information) for knowledge suggestion entries, an immutable "digital birth certificate" is established for each new piece of knowledge, ensuring that knowledge is traceable and verifiable from the moment it is discovered, providing a source guarantee for the trustworthy evolution of the knowledge base.

[0081] In some embodiments, before determining the target customer service dialogue data in response to a data processing instruction and generating real-time verification information based on the target customer service dialogue data, the method further includes: in response to acquiring the target customer service dialogue data at the initial acquisition time, converting the voice data in the target customer service dialogue data into initial text data; performing desensitization processing on the initial text data to obtain target text data; generating first verification information based on the target text data, and uploading the first verification information to the target blockchain for storage.

[0082] In the embodiments of this application, the initial acquisition time refers to the point in time when the customer service dialogue has just ended and the system automatically captures the complete original dialogue data; this is the starting point of the data lifecycle. Voice data refers to the original recording file or audio stream generated during the call between the customer service representative and the customer. Initial text data refers to the original dialogue text content obtained after converting the voice data using automatic speech recognition technology.

[0083] In embodiments of this application, in response to acquiring target customer service dialogue data at the initial acquisition time, the voice data in the target customer service dialogue data is converted into initial text data. The steps of this embodiment convert unstructured audio information, which is not easily analyzed directly, into structured text information that can be processed by machines.

[0084] In the embodiments of this application, desensitization refers to the process of using natural language processing technology to identify and mask or replace sensitive personal information (such as ID card number, bank card number, mobile phone number, etc.) in text.

[0085] In the embodiments of this application, target text data refers to clean text data obtained after desensitization processing, in which sensitive information has been removed or masked.

[0086] In the embodiments of this application, the initial text data is de-identified to obtain the target text data. The steps of this embodiment aim to strictly comply with data privacy protection regulations, eliminating sensitive personal information from the data before it is further used or stored as evidence, thereby ensuring privacy and security.

[0087] In the embodiments of this application, the first verification information refers to the data digest (such as a hash value) generated based on the clean target text data, which is the unique digital fingerprint of the initial state of the data.

[0088] In the embodiments of this application, first verification information is generated based on the target text data, and the first verification information is uploaded to the target blockchain for storage. The system performs a hash operation on the de-identified target text data to generate its unique digital fingerprint (first verification information), and immediately writes this fingerprint into the blockchain.

[0089] Through the embodiments of this application, by transcribing, de-identifying and completing the first blockchain notarization at the beginning of data generation, the authenticity, usability and privacy compliance of the data to be analyzed are ensured from the source, providing an immutable initial trust anchor for all subsequent processing and verification links.

[0090] In some embodiments, generating second verification information based on the quality inspection processing result includes: generating a target hash value based on the quality inspection processing result; obtaining target metadata corresponding to the quality inspection processing result, wherein the target metadata includes at least a processing timestamp; and generating second verification information based on the target hash value and the target metadata.

[0091] In the embodiments of this application, the target hash value refers to a fixed-length, unique digital digest obtained by calculating the complete content of the quality inspection processing result using a hash algorithm.

[0092] In the embodiments of this application, a target hash value is generated based on the quality inspection results. The system takes the quality inspection results as input and calculates them using a cryptographic hash function to generate a hash value that uniquely represents the content of the result.

[0093] In the embodiments of this application, target metadata refers to data describing the attributes related to the quality inspection processing results, including at least the time information of its generation. The processing timestamp refers to the specific point in time when the quality inspection processing results are analyzed and processed by the large language model.

[0094] In the embodiments of this application, the target metadata corresponding to the quality inspection processing result is obtained. The system needs to obtain key metadata such as the timestamp when the result was generated.

[0095] In the embodiments of this application, second verification information is generated based on the target hash value and target metadata. The steps of this embodiment involve combining the target hash value with its context information (target metadata) to form complete second verification information that can be stored and verified. For example, the two can be packaged into a single data structure, or the target metadata can also be hashed and uploaded to the blockchain along with the target hash value.

[0096] Through the embodiments of this application, verification information is generated by combining the content fingerprint of the quality inspection results with key time metadata, thereby enhancing the integrity and traceability of data storage and providing a more credible credential with richer context for subsequent verification and analysis of the authenticity of the results and their generation time.

[0097] In some embodiments, the method further includes: obtaining a target verification request, wherein the target verification request includes off-chain data to be verified; generating target verification information based on the off-chain data to be verified; obtaining on-chain verification information corresponding to the off-chain data to be verified, wherein the on-chain verification information includes first verification information or second verification information; and determining a target verification result based on the target verification information and the on-chain verification information.

[0098] In embodiments of this application, a target verification request refers to an instruction initiated by a user or system that requests verification of the integrity and authenticity of specific data. Off-chain data to be verified refers to data objects stored outside the blockchain that require verification. For example, off-chain data to be verified might be a text of a customer service conversation that claims to be the original, or a quality inspection report that claims to be the final one.

[0099] In the embodiments of this application, to obtain a target verification request, the system receives a clear verification instruction, which contains the specific data content that needs to be verified.

[0100] In the embodiments of this application, the target verification information refers to a data digest (e.g., a hash value) calculated in real time based on the currently provided off-chain data to be verified after receiving a verification request. It represents the state of the data at the moment of verification.

[0101] In the embodiments of this application, target verification information is generated based on the off-chain data to be verified. The system performs a hash operation on the off-chain data to be verified provided in the request to obtain its immediate digital digest.

[0102] In the embodiments of this application, on-chain verification information refers to the original data digest pre-stored in the target blockchain, corresponding to the off-chain data to be verified. Depending on the verification object, the on-chain verification information may be first verification information representing the original dialogue data, or second verification information representing the quality inspection result.

[0103] In the embodiments of this application, on-chain verification information corresponding to the off-chain data to be verified is obtained. Based on the context of the target verification request, the system locates and retrieves previously stored, corresponding on-chain verification information from the blockchain.

[0104] In the embodiments of this application, the target verification result refers to the conclusion drawn by comparing whether the target verification information is consistent with the on-chain verification information, which is used to clearly indicate whether the off-chain data to be verified is consistent with the original evidence version.

[0105] In the embodiments of this application, the target verification result is determined based on target verification information and on-chain verification information. The system precisely compares the target verification information calculated in real time with the on-chain verification information obtained from the blockchain. If they match, the verification passes; otherwise, the verification fails.

[0106] The embodiments of this application provide a standardized, readily initiable trusted verification mechanism, enabling any party to independently verify the authenticity and integrity of off-chain data without needing to trust the data provider. This greatly enhances the transparency and auditing capabilities of the entire system and provides a convenient tool for dispute resolution and compliance checks.

[0107] Figure 3 The schematic diagram illustrates the principle of a customer service dialogue data processing method according to an embodiment of this application.

[0108] like Figure 3 As shown, this embodiment describes a complete business process from the occurrence of a customer service dialogue to the updating of the knowledge base.

[0109] In the embodiments of this application, the data collection and initial evidence storage steps are performed first. A customer initiates an inquiry via telephone, engaging in a conversation with a bank agent. The customer service dialogue system automatically records the call. The system uses speech recognition technology to convert the call recording into text, and immediately performs dialogue transcription and anonymization processing on the text, automatically masking sensitive personal information such as the customer's name, ID number, and bank card number, generating clean text data. The anonymized text data is encrypted and stored on the bank's internal private server, i.e., off-chain storage and encryption. Simultaneously, the system calculates the hash value of the text data and adds metadata such as a timestamp and agent ID. The system packages the hash value and metadata generated in the previous step into an evidence storage transaction and submits it to the blockchain network. After consensus confirmation, a transaction hash is obtained, signifying that the "first verification information" of the original dialogue data has been permanently and immutably recorded on the blockchain.

[0110] Further, the system performs intelligent analysis and result notarization. When quality control of the conversation is required, the system, after authorization, allows the large language model to access anonymized text data stored off-chain. The large language model performs quality control analysis, specifically including: compliance checks, analyzing whether the conversation content violates regulatory rules such as illegal promises or misleading sales; service scoring, quantitatively scoring customer service quality from dimensions such as problem-solving efficiency, service attitude, and professionalism; and knowledge mining, identifying high-frequency, emerging, or uncollected user questions in the conversation and extracting high-quality answers provided by customer service. After completing the analysis, the large language model outputs a report and knowledge suggestions. The report includes compliance conclusions and service scores, while the knowledge suggestions are structured "question-answer" pairs. The system then calculates the hash value of the analysis results (including the report and knowledge suggestions) and performs blockchain notarization, storing the analysis results and adding "secondary verification information" representing the results of this intelligent analysis on the blockchain.

[0111] Further, the process involves review and knowledge base evolution. Knowledge suggestions are pushed to the review module, where knowledge base administrators manually review them to confirm their accuracy and applicability. Once approved by the administrator, the review operation (e.g., approval / rejection, reviewer, review time) is hashed as a new record and stored on the blockchain, forming the "third verification information" for the knowledge entry. This credible knowledge suggestion, which has undergone intelligent mining, manual review, and has all key steps stored on the blockchain, is officially imported into the knowledge base, completing the knowledge base update. Other customer service staff can then directly access this authoritative and credible standard answer from the knowledge base when encountering similar issues.

[0112] Through the embodiments of this application, by embedding blockchain evidence storage into key nodes of the business process (raw data, analysis results, audit records), the authenticity and traceability of data across the entire chain are ensured. Automated, multi-dimensional intelligent analysis is achieved through a large language model, significantly improving processing efficiency.

[0113] Based on the above-described customer service dialogue data processing method, this application also provides a customer service dialogue data processing device. The following will be combined with... Figure 4 The device is described in detail.

[0114] Figure 4 A schematic block diagram of a customer service dialogue data processing apparatus according to an embodiment of this application is shown.

[0115] like Figure 4 As shown, the customer service dialogue data processing device 400 of this embodiment includes a first processing module 410, a second processing module 420, a third processing module 430 and a fourth processing module 440.

[0116] The first processing module 410 is configured to respond to a data processing instruction, determine the target customer service dialogue data, and generate real-time verification information based on the target customer service dialogue data. The data processing instruction is used to trigger processing of the specified customer service dialogue data. In one embodiment, the first processing module 410 may be used to execute the operation S210 described above, which will not be repeated here.

[0117] The second processing module 420 is used to acquire the first verification information pre-stored in the target blockchain, and determine the first verification result based on the first verification information and real-time verification information. The first verification information is generated and stored in the target blockchain after the initial collection of the target customer service dialogue data. In one embodiment, the second processing module 420 can be used to execute the operation S220 described above, which will not be repeated here.

[0118] The third processing module 430 is used to process the target customer service dialogue data using a large language model in response to the first verification result indicating that the target customer service dialogue data has not been modified, and obtain the quality inspection processing result. In one embodiment, the third processing module 430 can be used to perform the operation S230 described above, which will not be repeated here.

[0119] The fourth processing module 440 is used to generate second verification information based on the quality inspection results, and upload the second verification information to the target blockchain for storage. In one embodiment, the fourth processing module 440 can be used to perform the operation S240 described above, which will not be repeated here.

[0120] According to embodiments of this application, any plurality of modules among the first processing module 410, the second processing module 420, the third processing module 430, and the fourth processing module 440 may be combined into one module, or any one of these modules may be split into multiple modules. Alternatively, at least a portion of the functionality of one or more of these modules may be combined with at least a portion of the functionality of other modules and implemented in one module. According to embodiments of this application, at least one of the first processing module 410, the second processing module 420, the third processing module 430, and the fourth processing module 440 may be at least partially implemented as hardware circuitry, such as a field-programmable gate array (FPGA), a programmable logic array (PLA), a system-on-a-chip, a system-on-a-substrate, a system-on-package, an application-specific integrated circuit (ASIC), or implemented in hardware or firmware by any other reasonable means of integrating or packaging the circuitry, or implemented in any one of the three implementation methods of software, hardware, and firmware, or in a suitable combination of any of these. Alternatively, at least one of the first processing module 410, the second processing module 420, the third processing module 430, and the fourth processing module 440 may be implemented at least partially as a computer program module, which can perform corresponding functions when the computer program module is run.

[0121] In some embodiments, the third processing module includes: a quality inspection submodule, used to perform compliance quality inspection on the target customer service dialogue data to obtain a violation statement identification result, wherein the violation statement identification result indicates whether the dialogue content corresponding to the dialogue data complies with predefined regulatory rules; and an evaluation submodule, used to perform service quality evaluation on the target customer service dialogue data to obtain a service quality score, wherein the service quality score; wherein the quality inspection processing result includes the violation statement identification result and the service quality score.

[0122] In some embodiments, the third processing module further includes: a determination submodule, configured to determine target dialogue data from the target customer service dialogue data based on the violation statement identification result and the service quality score, wherein the target dialogue data is a content fragment in the target customer service dialogue data that does not contain violation statements and whose corresponding service quality score is higher than a preset threshold; and a first generation submodule, configured to generate knowledge suggestion entries based on the target dialogue data, wherein the knowledge suggestion entries include a question description and an answer description.

[0123] In some embodiments, the apparatus further includes: a first generation module, configured to generate third verification information based on knowledge suggestion entries after processing the target customer service dialogue data using a large language model to obtain a quality inspection result in response to a first verification result indicating that the target customer service dialogue data has not been modified; and an upload module, configured to upload the third verification information to the target blockchain for storage.

[0124] In some embodiments, the apparatus further includes: a conversion module, configured to convert voice data in the target customer service dialogue data into initial text data in response to acquiring the target customer service dialogue data at the initial acquisition time, before determining the target customer service dialogue data in response to a data processing instruction and generating real-time verification information based on the target customer service dialogue data; a first processing module, configured to perform desensitization processing on the initial text data to obtain target text data; and a second processing module, configured to generate first verification information based on the target text data and upload the first verification information to the target blockchain for storage.

[0125] In some embodiments, the fourth processing module includes: a second generation submodule, used to generate a target hash value based on the quality inspection processing result; an acquisition submodule, used to acquire target metadata corresponding to the quality inspection processing result, wherein the target metadata includes at least a processing timestamp; and a third generation submodule, used to generate second verification information based on the target hash value and the target metadata.

[0126] In some embodiments, the apparatus further includes: a first acquisition module, configured to acquire a target verification request, wherein the target verification request includes off-chain data to be verified; a second generation module, configured to generate target verification information based on the off-chain data to be verified; a second acquisition module, configured to acquire on-chain verification information corresponding to the off-chain data to be verified, wherein the on-chain verification information includes first verification information or second verification information; and a determination module, configured to determine a target verification result based on the target verification information and the on-chain verification information.

[0127] Figure 5 A block diagram schematically illustrates an electronic device suitable for implementing a customer service dialogue data processing method according to an embodiment of this application.

[0128] like Figure 5 As shown, an electronic device 500 according to an embodiment of this application includes a processor 501, which can perform various appropriate actions and processes according to a program stored in a read-only memory (ROM) 502 or a program loaded from a storage portion 508 into a random access memory (RAM) 503. The processor 501 may include, for example, a general-purpose microprocessor (e.g., a CPU), an instruction set processor and / or an associated chipset and / or a special-purpose microprocessor (e.g., an application-specific integrated circuit (ASIC)), etc. The processor 501 may also include onboard memory for caching purposes. The processor 501 may include a single processing unit or multiple processing units for performing different actions of the method flow according to an embodiment of this application.

[0129] RAM 503 stores various programs and data required for the operation of electronic device 500. Processor 501, ROM 502, and RAM 503 are interconnected via bus 504. Processor 501 executes various operations of the method flow according to embodiments of this application by executing programs in ROM 502 and / or RAM 503. It should be noted that programs may also be stored in one or more memories other than ROM 502 and RAM 503. Processor 501 may also execute various operations of the method flow according to embodiments of this application by executing programs stored in one or more memories.

[0130] According to embodiments of this application, the electronic device 500 may further include an input / output (I / O) interface 505, which is also connected to a bus 504. The electronic device 500 may also include one or more of the following components connected to the input / output (I / O) interface 505: an input section 506 including a keyboard, mouse, etc.; an output section 507 including a cathode ray tube (CRT), liquid crystal display (LCD), etc., and a speaker, etc.; a storage section 508 including a hard disk, etc.; and a communication section 509 including a network interface card such as a LAN card, modem, etc. The communication section 509 performs communication processing via a network such as the Internet. A drive 510 is also connected to the input / output (I / O) interface 505 as needed. A removable medium 511, such as a disk, optical disk, magneto-optical disk, semiconductor memory, etc., is installed on the drive 510 as needed so that computer programs read from it can be installed into the storage section 508 as needed.

[0131] This application also provides a computer-readable storage medium, which may be included in the device / apparatus / system described in the above embodiments; or it may exist independently and not assembled into the device / apparatus / system. The computer-readable storage medium carries one or more programs, which, when executed, implement the method according to the embodiments of this application.

[0132] According to embodiments of this application, the computer-readable storage medium can be a non-volatile computer-readable storage medium, such as including but not limited to: portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination thereof. In this application, the computer-readable storage medium can be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system, apparatus, or device. For example, according to embodiments of this application, the computer-readable storage medium may include ROM 502 and / or RAM 503 and / or one or more memories other than ROM 502 and RAM 503 described above.

[0133] Embodiments of this application also include a computer program product comprising a computer program containing program code for performing the methods shown in the flowchart. When the computer program product is run on a computer system, the program code enables the computer system to implement the customer service dialogue data processing method provided in the embodiments of this application.

[0134] When the computer program is executed by the processor 501, it performs the functions defined in the system / apparatus of this application embodiment. According to the embodiments of this application, the systems, apparatuses, modules, units, etc., described above can be implemented by computer program modules.

[0135] In one embodiment, the computer program may rely on a tangible storage medium such as an optical storage device or a magnetic storage device. In another embodiment, the computer program may also be transmitted and distributed in the form of signals over a network medium, and may be downloaded and installed via the communication section 509, and / or installed from a removable medium 511. The program code contained in the computer program can be transmitted using any suitable network medium, including but not limited to: wireless, wired, etc., or any suitable combination thereof.

[0136] In such an embodiment, the computer program can be downloaded and installed from a network via communication section 509, and / or installed from removable medium 511. When the computer program is executed by processor 501, it performs the functions defined in the system of this application embodiment. According to embodiments of this application, the systems, devices, apparatuses, modules, units, etc., described above can be implemented by computer program modules.

[0137] According to embodiments of this application, program code for executing the computer programs provided in the embodiments of this application can be written in any combination of one or more programming languages. Specifically, these computational programs can be implemented using high-level procedural and / or object-oriented programming languages, and / or assembly / machine languages. The program code can be executed entirely on the user's computing device, partially on the user's device, partially on a remote computing device, or entirely on a remote computing device or server. In cases involving remote computing devices, the remote computing device can be connected to the user's computing device via any type of network, including a local area network (LAN) or a wide area network (WAN), or it can be connected to an external computing device (e.g., via the Internet using an Internet service provider).

[0138] The flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer program products according to various embodiments of this application. In this regard, each block in a flowchart or block diagram may represent a module, segment, or portion of code containing one or more executable instructions for implementing a specified logical function. It should also be noted that in some alternative implementations, the functions indicated in the blocks may occur in a different order than those indicated in the drawings. For example, two consecutively indicated blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in a block diagram or flowchart, and combinations of blocks in a block diagram or flowchart, may be implemented using a dedicated hardware-based system that performs the specified function or operation, or using a combination of dedicated hardware and computer instructions.

[0139] Those skilled in the art will understand that the features described in the various embodiments of this application can be combined and / or combined in various ways, even if such combinations or combinations are not explicitly described in this application. In particular, the features described in the various embodiments of this application can be combined and / or combined in various ways without departing from the spirit and teachings of this application. All such combinations and / or combinations fall within the scope of this application.

Claims

1. A method for processing customer service dialogue data, characterized in that, The method includes: In response to a data processing instruction, the target customer service dialogue data is determined, and real-time verification information is generated based on the target customer service dialogue data, wherein the data processing instruction is used to trigger the processing of the specified customer service dialogue data; Obtain the first verification information pre-stored in the target blockchain, and determine the first verification result based on the first verification information and the real-time verification information, wherein the first verification information is the verification information generated and stored in the target blockchain after the initial collection of the target customer service dialogue data is completed; In response to the first verification result indicating that the target customer service dialogue data has not been modified, the target customer service dialogue data is processed using a large language model to obtain the quality inspection result; Based on the quality inspection results, second verification information is generated and uploaded to the target blockchain for storage.

2. The method according to claim 1, characterized in that, The process of using a large language model to process the target customer service dialogue data to obtain quality inspection results includes: The target customer service dialogue data is subjected to compliance quality inspection to obtain the violation expression identification result, wherein the violation expression identification result indicates whether the dialogue content corresponding to the dialogue data complies with predefined regulatory rules; The target customer service dialogue data is used to evaluate service quality and obtain a service quality score, wherein the service quality score is... The quality inspection results include the identification results of the non-compliant statements and the service quality score.

3. The method according to claim 2, characterized in that, The process of using a large language model to process the target customer service dialogue data to obtain quality inspection results also includes: Based on the violation statement identification results and the service quality score, target dialogue data is determined from the target customer service dialogue data, wherein the target dialogue data is a content segment in the target customer service dialogue data that does not contain violation statements and whose corresponding service quality score is higher than a preset threshold. Based on the target dialogue data, knowledge suggestion entries are generated, wherein the knowledge suggestion entries include a question description and an answer description.

4. The method according to claim 3, characterized in that, After the method responds to the first verification result indicating that the target customer service dialogue data has not been modified, processes the target customer service dialogue data using a large language model, and obtains a quality inspection result, the method further includes: Based on the knowledge suggestion entries, third verification information is generated; The third verification information is uploaded to the target blockchain for storage.

5. The method according to claim 1, characterized in that, Before responding to a data processing instruction, determining the target customer service dialogue data, and generating real-time verification information based on the target customer service dialogue data, the method further includes: In response to acquiring the target customer service dialogue data at the initial acquisition time, the voice data in the target customer service dialogue data is converted into initial text data; The initial text data is anonymized to obtain the target text data; Based on the target text data, the first verification information is generated and uploaded to the target blockchain for storage.

6. The method according to claim 1, characterized in that, The generation of second verification information based on the quality inspection results includes: Based on the quality inspection results, a target hash value is generated; Obtain the target metadata corresponding to the quality inspection processing result, wherein the target metadata includes at least a processing timestamp; The second verification information is generated based on the target hash value and the target metadata.

7. The method according to any one of claims 1 to 6, characterized in that, The method further includes: Obtain a target verification request, wherein the target verification request includes off-chain data to be verified; Based on the off-chain data to be verified, generate target verification information; Obtain on-chain verification information corresponding to the off-chain data to be verified, wherein the on-chain verification information includes the first verification information or the second verification information; Based on the target verification information and the on-chain verification information, the target verification result is determined.

8. A customer service dialogue data processing device, characterized in that, The device includes: The first processing module is used to respond to a data processing instruction, determine the target customer service dialogue data, and generate real-time verification information based on the target customer service dialogue data, wherein the data processing instruction is used to trigger the processing of the specified customer service dialogue data. The second processing module is used to obtain the first verification information pre-stored in the target blockchain, and determine the first verification result based on the first verification information and the real-time verification information, wherein the first verification information is the verification information generated and stored in the target blockchain after the target customer service dialogue data is initially collected; The third processing module is used to process the target customer service dialogue data using a large language model in response to the first verification result indicating that the target customer service dialogue data has not been modified, and to obtain the quality inspection processing result. The fourth processing module is used to generate second verification information based on the quality inspection results, and upload the second verification information to the target blockchain for storage.

9. An electronic device, comprising: One or more processors; Memory, used to store one or more computer programs. The characteristic feature is that the one or more processors execute the one or more computer programs to implement the steps of the method according to any one of claims 1 to 7.

10. A computer-readable storage medium having a computer program or instructions stored thereon, characterized in that, When the computer program or instructions are executed by a processor, they implement the steps of the method according to any one of claims 1 to 7.

11. A computer program product, comprising a computer program or instructions, characterized in that, When the computer program or instructions are executed by a processor, they implement the steps of the method according to any one of claims 1 to 7.

Images