Government affair service hotline processing method and system
By combining high-precision speech recognition and large language models, calls to the government service hotline are automatically processed, solving the problem of insufficient reception staff, improving the efficiency and quality of the government service hotline, and realizing intelligent management and data support of the dialogue process.
Patent Information
- Application Number
- CN202510744652.4
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-06-05
- Publication Date
- 2025-09-12
AI Technical Summary
Traditional government service hotline systems face insufficient switchboard staffing, difficulty in efficiently responding to large-scale and highly complex calls, and lack a systematic call content management mechanism.
High-precision speech recognition technology and large language models, combined with intelligent agent technology, enable automated processing of conversational processes, including speech-to-text conversion, semantic analysis, key information extraction, knowledge base retrieval, and disposal opinion recommendations, as well as work order generation, data mining, and trend analysis.
Significantly improve the work efficiency and service quality of government service hotlines, reduce the burden on manual seats, improve response speed and quality, provide data support to assist decision-making, and achieve process transparency and traceability.
Smart Images
Figure CN120636385A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the field of information technology, and in particular to a method and system for handling government service hotlines. Background Art
[0002] Traditional government service hotline systems face multiple challenges, including insufficient operator staff, difficulty in efficiently responding to large-scale and highly complex calls, and a lack of a systematic call content management mechanism.
[0003] In recent years, the rise of large language model technology has provided technical support for solving these problems. These models, with their ability to understand and generate natural language, can automate conversational processes. Furthermore, the integration of intelligent agent technology enables real-time monitoring and analysis of data dynamics, providing proactive early warning support for abnormal situations. Summary of the Invention
[0004] The purpose of the present invention is to provide a method and system for handling government service hotlines to solve the problems raised in the above background technology.
[0005] To achieve the above-mentioned purpose, the present invention provides the following technical solution: a method for handling government service hotlines, comprising the following steps of hotline information processing:
[0006] Leveraging high-precision speech recognition technology, the system converts user speech into text, identifies voice roles, and uses natural language processing algorithms for semantic analysis. It then categorizes issues into corresponding municipal service sections based on pre-set classification standards, tags, extracts key information, and generates work order information.
[0007] The converted text information is input into the large language model for preliminary content summary;
[0008] We conduct an in-depth analysis of the summarized content and extract keywords and phrases directly related to the disposal work, covering important dimensions such as specific policy clauses, geographic location information, and time information.
[0009] Preferably, the speech recognition process in hotline information processing uses advanced acoustic models and language models to ensure accurate transcription of call content in various background noise environments, accurate distinction between callers and answering personnel, and real-time transmission of converted text information to the backend database.
[0010] Preferably, the method further includes a treatment process, wherein the treatment process includes the following steps:
[0011] Knowledge base construction: Enter relevant policy information into the knowledge base to establish a knowledge base containing policies, regulations, and frequently asked questions, and the knowledge base can be dynamically updated;
[0012] Recommended solutions: Based on the extracted key information, the knowledge base is searched for matching answers or solutions. Simultaneously, a large language model performs a secondary evaluation of the search results, considering the logic and completeness of the answers, and the applicability and effectiveness of the solutions, to provide more personalized and accurate solutions.
[0013] Generate a handling work order and store the handling results: Based on the finalized handling opinions and the extracted key information, a handling work order containing caller details, problem description, and handling suggestions is automatically generated, and all relevant information is stored in the database.
[0014] Preferably, the method further includes summarizing the result data, and the summarizing the result data includes the following steps:
[0015] The model performs a secondary summary based on the work order content: it generates tags for processed work orders using a large language model, and conducts data mining and trend analysis on the work order data by merging similar tags and performing multi-level tag convergence to identify potential issues and service improvement points.
[0016] Summarize data trends and provide early warning information: Based on in-depth mining of massive work order data, predict potential service pain points and sudden risks in specific areas or groups, and push warning information to responsible departments through an intelligent early warning mechanism.
[0017] Preferably, through the steps of hotline information processing, handling process and result data aggregation, the working efficiency and service quality of the government service hotline can be significantly improved, providing data support for city managers to assist in decision-making.
[0018] A system for handling government service hotlines includes a hotline information processing module: used to receive voice messages from users calling a government service hotline, convert the voice messages into text using high-precision voice recognition technology, identify voice roles, perform semantic analysis using a natural language processing algorithm, classify issues into corresponding municipal service sections based on preset classification standards, mark and extract key information to generate work order information; at the same time, input the converted text information into a large language model for preliminary content summary, and further extract keywords and phrases directly related to the handling work.
[0019] Preferably, the speech recognition submodule in the hotline information processing module uses advanced acoustic models and language models to ensure accurate transcription of incoming call content in various background noise environments, accurately distinguish between callers and answering personnel, and transmit the converted text information to the background database in real time to ensure accurate information transmission and lay the foundation for subsequent intelligent analysis and processing.
[0020] Preferably, a treatment process module is also included, and the treatment process module includes the following submodules:
[0021] Knowledge base construction submodule: used to enter relevant policy information into the knowledge base, and establish a knowledge base that includes policies, regulations, and FAQs and can be dynamically updated;
[0022] The action suggestion submodule is used to search for matching answers or solutions in the knowledge base based on the extracted key information. At the same time, the large language model conducts a secondary evaluation of the search results, considering the logic and completeness of the answers, and the applicability and effectiveness of the solutions, to provide more personalized and accurate action suggestions.
[0023] The sub-module for generating handling work orders and storing handling results is used to automatically generate handling work orders containing caller details, problem description, and handling suggestions based on the finalized handling opinions and extracted key information, and store all relevant information in the database.
[0024] Preferably, a result data summary module is also included, and the result data summary module includes the following submodules:
[0025] Model secondary summary submodule: This module generates labels for processed work orders using a large language model. It also conducts data mining and trend analysis on work order data by merging similar labels and performing multi-level label convergence to identify potential issues and service improvement points.
[0026] Data trend summary and early warning sub-module: used to conduct in-depth mining based on massive work order data, predict potential service pain points and sudden risks in specific areas or groups, and push warning information to responsible departments through an intelligent early warning mechanism.
[0027] Preferably, through the collaborative work of the hotline information processing module, the handling process module and the result data summary module, the working efficiency and service quality of the government service hotline are significantly improved, and data support is provided to city managers to assist in decision-making. With the help of a complete work order generation and data storage mechanism, full tracking and management of each incoming call is achieved to ensure process transparency and traceability, while providing strong support for model service tuning and training, and improving problem identification and processing capabilities.
[0028] Compared with the prior art, the present invention has the following beneficial effects:
[0029] The method and system for handling government service hotlines proposed in the present invention realize the automated processing of incoming call content through intelligent means, thereby effectively reducing the degree of manpower utilization. At the same time, by conducting a secondary summary and analysis of historical work order data, the data trends and potential early warning information therein are accurately analyzed, thereby achieving the goal of improving service quality. The automation of the government service hotline processing process is realized. This innovation effectively reduces the workload of manual agents and significantly improves the service response speed and quality. The system has the ability to deeply mine and analyze massive amounts of data, can accurately identify potential problems and issue early warnings in a timely manner, and provide strong data support for management to assist them in making more scientific and reasonable decisions. BRIEF DESCRIPTION OF THE DRAWINGS
[0030] Figure 1 Flow chart of the method of the present invention. DETAILED DESCRIPTION
[0031] In order to clearly and completely describe the objectives and technical solutions of the present invention and make the advantages more clearly understood, the embodiments of the present invention are further described in detail below with reference to the accompanying drawings. It should be understood that the specific embodiments described herein are part of the embodiments of the present invention, not all of them, and are only used to explain the embodiments of the present invention, not to limit the embodiments of the present invention. All other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.
[0032] For example 1, please refer to Figure 1 The present invention provides a technical solution: a method for handling a government service hotline, comprising the following steps:
[0033] S1. Hotline information processing
[0034] S101. Speech Recognition: When users call the government service hotline, the system uses high-precision speech recognition technology to convert their voice into text. This process utilizes advanced acoustic and language models to ensure accurate transcription of call content in all types of background noise. Even in noisy environments or with accents, the system effectively recognizes the voice, ensuring accurate transmission of information. It also identifies voice roles, accurately distinguishing between callers and answering personnel, laying the foundation for subsequent intelligent analysis and processing. The converted text is transmitted to a backend database in real time, where powerful natural language processing algorithms perform semantic analysis to deeply identify the core points of the user's request. Based on pre-set classification criteria, issues are automatically categorized into corresponding municipal service areas, such as municipal facility maintenance, environmental sanitation complaints, and public resource allocation, allowing for rapid matching to the appropriate processing procedures. The system also tags and extracts key information, such as the specific location involved, event characteristics, and urgency, generating detailed work order information. This accurate data allows relevant departments to respond quickly and efficiently to user feedback, improving the accuracy and timeliness of hotline services and enhancing public satisfaction and trust in government services.
[0035] S102. Content Summarization: The converted text message is fed into the Big Language Model for preliminary content summarization. The Big Language Model understands natural language and, based on the context, provides a concise and comprehensive summary of the call content, helping to quickly identify the caller's primary concern or question. This allows the core content of the call to be grasped quickly, providing strong support for subsequent processing.
[0036] S103. Extracting Key Information: When conducting an in-depth analysis of the summarized content, it is necessary to further extract keywords and phrases directly related to the disposal work. These keywords and phrases involve multiple important dimensions, such as specific policy clauses, which may be related to policy provisions in different fields and at different levels, covering specific content such as the policy's scope of application, implementation standards, and preferential conditions. In terms of geographic location information, the specific location and regional scope should be clearly defined, whether it is a specific city, township, or precise geographic coordinates, as well as the unique characteristics of the geographic location and related influencing factors. Time information is also crucial, including the policy's effective date, the disposal task deadline, and key time nodes at each stage.
[0037] This critical information plays a crucial role in subsequent processing steps and is an indispensable basis. Accurately extracting specific policy terms ensures that disposal work adheres strictly to relevant regulations and prevents violations. Clarifying geographic location information facilitates the rational planning of resource allocation and routes of action, improving disposal efficiency. Accurately grasping time information ensures that all disposal tasks are completed on time, advancing the disposal process in an orderly manner.
[0038] S2. Disposal process
[0039] S201. Knowledge Base Construction: Relevant policy information is stored in the knowledge base, creating a comprehensive knowledge base encompassing policies, regulations, and frequently asked questions. This knowledge base not only includes static information resources but also dynamically updates to adapt to the latest policy changes and social needs. By continuously enriching and improving the knowledge base, the system ensures that when faced with complex issues, it can provide the most up-to-date and accurate answers and solutions.
[0040] S202, Recommended Actions: After extracting key information, the system will quickly and accurately search the knowledge base. Based on the characteristics of the key information, the system will carefully screen numerous knowledge items to accurately match the corresponding answer or feasible solution.
[0041] At the same time, the large language model performs a secondary evaluation of the search results. Leveraging its powerful language understanding and analysis capabilities, the large language model is able to deeply analyze the content of search results. It considers multiple dimensions, such as the logic and completeness of the answer, and the applicability and effectiveness of the solution. Through this in-depth evaluation, the large language model can provide more personalized and precise recommendations based on the user's specific needs and context.
[0042] This dual verification mechanism, combining knowledge base retrieval with large language model evaluation, provides two solid lines of defense for information accuracy and reliability. On the one hand, knowledge base retrieval ensures the accuracy and relevance of basic information; on the other hand, large language model evaluation further optimizes and refines the results, making them more tailored to users' actual needs. This mechanism effectively improves the accuracy and reliability of response advice, ensuring that users seeking help receive satisfactory responses and effectively resolve their problems.
[0043] S203. Generate a work order and store the results: Based on the finalized resolution and extracted key information, the system automatically and efficiently generates a corresponding work order. This work order is comprehensive and detailed, including caller details such as name, contact information, and location, facilitating effective communication and follow-up. The problem description details the specific issue raised by the caller, including the specific time, scenario, and possible related factors, striving to clearly present the details of the issue. Furthermore, the work order also fully records the resolution suggestions, clearly defining the specific measures to be taken, the order of priority, and key considerations.
[0044] All relevant information is stored in the database, which not only facilitates subsequent inquiries, whether staff need to quickly find specific call information or conduct comprehensive data query and analysis, they can quickly and accurately obtain the required data; it also lays a solid foundation for subsequent data analysis. Through in-depth mining and analysis of the data, potential patterns and problems can be discovered, providing a strong basis for decision-making.
[0045] With a comprehensive work order generation and data storage mechanism, every incoming call can be tracked and managed throughout its entire lifecycle. From call reception to issue resolution and final outcome, every step is recorded and tracked in detail, ensuring process transparency and traceability. Furthermore, rich and accurate data provides strong support for model service tuning and training. By learning and analyzing large amounts of real-world data, the model can continuously optimize its algorithms and parameters, improving its problem identification and resolution capabilities, thereby providing users with more accurate and efficient services.
[0046] S3. Result data summary
[0047] S301: The model performs a secondary summary based on the work order content. All processed work orders are labeled using a large language model. Through similar label merging and multi-level label convergence, the work order data is subjected to deeper data mining and trend analysis. This process helps identify potential issues and service improvements. Through in-depth analysis of large amounts of work order data, common issues and changing trends can be discovered, providing a basis for optimizing service processes.
[0048] S302. Data Trend Aggregation and Early Warning Information: Based on in-depth mining of massive amounts of work order data, the system accurately predicts potential service pain points and unexpected risks in specific areas or groups. Through an intelligent early warning mechanism, it proactively sends warnings to responsible departments, enabling them to quickly respond and develop countermeasures. This forward-looking early warning system effectively activates resource pre-deployment capabilities, eliminates potential risks before they arise, and significantly improves service response efficiency and quality control.
[0049] This technical solution significantly improves the efficiency and service quality of the government service hotline, while providing strong data support for city managers, helping them make more scientific and reasonable decisions.
[0050] Example 2, based on Example 1, proposes a system for a method of handling government service hotlines, including a hotline information processing module: for receiving voice messages from users calling a government service hotline, converting the voice messages into text with the help of high-precision voice recognition technology, identifying voice roles, performing semantic analysis using natural language processing algorithms, classifying problems into corresponding municipal service sections according to preset classification standards, marking and extracting key information to generate work order information; at the same time, inputting the converted text information into a large language model for preliminary content summary, and further extracting keywords and phrases directly related to the handling work.
[0051] The speech recognition submodule in the hotline information processing module uses advanced acoustic and language models to ensure accurate transcription of incoming call content in various background noise environments, accurately distinguish between callers and answerers, and transmit the converted text information to the backend database in real time to ensure accurate information transmission and lay the foundation for subsequent intelligent analysis and processing.
[0052] It also includes a handling process module, which includes the following sub-modules: a knowledge base construction sub-module: used to enter relevant policy information into the knowledge base, and establish a knowledge base that contains policies, regulations, and frequently asked questions and can be dynamically updated; a handling opinion recommendation sub-module: used to retrieve matching answers or solutions in the knowledge base based on the extracted key information, and at the same time, the large language model conducts a secondary evaluation of the retrieval results, considering the logic and completeness of the answers, the applicability and effectiveness of the solutions, and providing more personalized and accurate handling suggestions; a handling work order generation and handling result storage sub-module: used to automatically generate a handling work order containing caller details, problem description, and handling suggestions based on the finalized handling opinions and the extracted key information, and store all relevant information in the database.
[0053] It also includes a result data summary module, which includes the following sub-modules: Model secondary summary sub-module: used to generate labels for processed work orders through a large language model, and conduct data mining and trend analysis on work order data through similar label merging and multi-level label convergence to identify potential problems and service improvement points; Data trend summary and early warning sub-module: used to deeply mine massive work order data, predict potential service pain points and sudden risks in specific areas or groups, and push warning information to responsible departments through intelligent early warning mechanisms.
[0054] Through the collaborative work of the hotline information processing module, the handling process module and the result data aggregation module, the working efficiency and service quality of the government service hotline are significantly improved, providing data support for city managers to assist in decision-making, and with the help of a complete work order generation and data storage mechanism, full tracking and management of each incoming call is achieved to ensure process transparency and traceability. At the same time, it provides strong support for model service tuning and training, and improves problem identification and handling capabilities.
[0055] While embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions, and variations may be made to these embodiments without departing from the principles and spirit of the invention, and that the scope of the invention is defined by the appended claims and their equivalents.
Claims
1. A method for handling government service hotlines, characterized by: Including hotline information processing steps: Leveraging high-precision speech recognition technology, the system converts user speech into text, identifies voice roles, and uses natural language processing algorithms for semantic analysis. It then categorizes issues into corresponding municipal service sections based on pre-set classification standards, tags, extracts key information, and generates work order information. The converted text information is input into the large language model for preliminary content summary; We conduct an in-depth analysis of the summarized content and extract keywords and phrases directly related to the disposal work, covering important dimensions such as specific policy clauses, geographic location information, and time information.
2. A method for handling government service hotlines according to claim 1, characterized in that: The speech recognition process in hotline information processing uses advanced acoustic models and language models to ensure accurate transcription of incoming call content in various background noise environments, accurately distinguish between callers and answering personnel, and transmit the converted text information to the backend database in real time.
3. The method for handling government service hotline according to claim 2, characterized in that: It also includes a disposal process, which includes the following steps: Knowledge base construction: Enter relevant policy information into the knowledge base to establish a knowledge base containing policies, regulations, and frequently asked questions, and the knowledge base can be dynamically updated; Recommended solutions: Based on the extracted key information, the knowledge base is searched for matching answers or solutions. Simultaneously, a large language model performs a secondary evaluation of the search results, considering the logic and completeness of the answers, and the applicability and effectiveness of the solutions, to provide more personalized and accurate solutions. Generate a handling work order and store the handling results: Based on the finalized handling opinions and the extracted key information, a handling work order containing caller details, problem description, and handling suggestions is automatically generated, and all relevant information is stored in the database.
4. A method for handling government service hotlines according to claim 3, characterized in that: It also includes result data summary, which includes the following steps: The model performs a secondary summary based on the work order content: it generates tags for processed work orders using a large language model, and conducts data mining and trend analysis on the work order data by merging similar tags and performing multi-level tag convergence to identify potential issues and service improvement points. Summarize data trends and provide early warning information: Based on in-depth mining of massive work order data, predict potential service pain points and sudden risks in specific areas or groups, and push warning information to responsible departments through an intelligent early warning mechanism.
5. A method for handling government service hotline according to claim 4, characterized in that: Through the steps of hotline information processing, handling process and result data aggregation, the efficiency and service quality of the government service hotline can be significantly improved, providing data support for city managers to assist in decision-making.
6. A system for the government service hotline handling method according to claim 5, characterized in that: It includes a hotline information processing module: used to receive voice messages from users calling the government service hotline, convert the voice messages into text with the help of high-precision speech recognition technology, identify voice roles, perform semantic analysis using natural language processing algorithms, classify problems into corresponding municipal service sections according to preset classification standards, mark and extract key information to generate work order information; at the same time, the converted text information is input into a large language model for preliminary content summary, and further extract keywords and phrases directly related to the handling work.
7. A system according to claim 6, characterized in that: The speech recognition submodule in the hotline information processing module uses advanced acoustic and language models to ensure accurate transcription of incoming call content in various background noise environments, accurately distinguish between callers and answerers, and transmit the converted text information to the backend database in real time to ensure accurate information transmission and lay the foundation for subsequent intelligent analysis and processing.
8. A system according to claim 6, characterized in that: It also includes a disposal process module, which includes the following submodules: Knowledge base construction submodule: used to enter relevant policy information into the knowledge base, and establish a knowledge base that includes policies, regulations, and FAQs and can be dynamically updated; The action suggestion submodule is used to search for matching answers or solutions in the knowledge base based on the extracted key information. At the same time, the large language model conducts a secondary evaluation of the search results, considering the logic and completeness of the answers, and the applicability and effectiveness of the solutions, to provide more personalized and accurate action suggestions. The sub-module for generating handling work orders and storing handling results is used to automatically generate handling work orders containing caller details, problem description, and handling suggestions based on the finalized handling opinions and extracted key information, and store all relevant information in the database.
9. A system according to claim 6, characterized in that: It also includes a result data summary module, which includes the following submodules: Model secondary summary submodule: This module generates labels for processed work orders using a large language model. It also conducts data mining and trend analysis on work order data by merging similar labels and performing multi-level label convergence to identify potential issues and service improvement points. Data trend summary and early warning sub-module: used to conduct in-depth mining based on massive work order data, predict potential service pain points and sudden risks in specific areas or groups, and push warning information to responsible departments through an intelligent early warning mechanism.
10. A system according to claim 6, characterized in that: Through the collaborative work of the hotline information processing module, the handling process module and the result data aggregation module, the working efficiency and service quality of the government service hotline are significantly improved, providing data support for city managers to assist in decision-making, and with the help of a complete work order generation and data storage mechanism, full tracking and management of each incoming call is achieved to ensure process transparency and traceability. At the same time, it provides strong support for model service tuning and training, and improves problem identification and handling capabilities.