Conversational structured information acquisition method and system

Abstract

The invention discloses a dialogue type structured information acquisition method and system. The method comprises the following steps: receiving a dialogue request initiated by a user; according to the service type of the dialogue request, collecting answer content fed back by a user based on a preset intelligent template; judging whether the answer content meets a preset answer requirement or not; if yes, judging whether all dialogue contents contained in the intelligent template are collected or not; if yes, obtaining all dialogue contents, and extracting all answer contents from all dialogue contents according to a preset rule; and carrying out format conversion on all the answer contents to obtain target answer information, and storing the target answer information through an HTTP (Hyper Text Transport Protocol) interface. Therefore, through the flexible guidance of the intelligent agent and the dynamic dialogue strategy, the user naturally interacts with the intelligent assistant in the information collection process, the participation degree and the satisfaction degree are improved, the abandoning situation caused by poor experience is reduced, and then the traditional information collection mode is effectively changed.

Description

Technical Field

[0001] The present application relates to the field of artificial intelligence and data acquisition technology, and in particular to a method and system for collecting conversational structured information. Background Art

[0002] With the advent of the information age and the development of big data and intelligent technologies, the demand for data collection is increasing. Data collection is a fundamental task in the daily work of enterprises, governments, and various institutions, and is widely used in various fields such as surveys, market analysis, user behavior research, and financial auditing.

[0003] Traditional data collection methods usually rely on manual filling out of forms or simple form-based input interfaces. Although these methods are feasible in some simple application scenarios, they usually require users to understand and fill in the field content on their own. For complex or changeable information collection scenarios, users may feel confused due to lack of guidance and even give up filling in the form. Once the form is set up, users often find it difficult to flexibly respond to the needs of different situations, which reduces the user experience.

[0004] Therefore, there is an urgent need to improve the data collection process through intelligent and automated technologies. Summary of the Invention

[0005] Based on the above-mentioned deficiencies of the prior art, the present application provides a method and system for collecting conversational structured information to solve the problem of poor user experience brought about by the prior art.

[0006] In order to achieve the above objectives, this application provides the following technical solutions:

[0007] In a first aspect, the present application provides a method for collecting conversational structured information, comprising:

[0008] Receive conversation requests initiated by users;

[0009] According to the business type of the dialogue request, collecting the user's response content based on the preset intelligent template;

[0010] Determine whether the answer content meets the preset answer requirements;

[0011] If the answer content meets the preset answer requirements, then determine whether all the conversation content included in the smart template has been collected;

[0012] If all the conversation contents included in the smart template have been collected, all the conversation contents are obtained, and all the answer contents are extracted from the conversation contents according to the preset rules;

[0013] All the answer contents are formatted to obtain target answer information, and the target answer information is stored through an HTTP interface.

[0014] Optionally, in the above-mentioned method for collecting conversational structured information, the method for obtaining the preset intelligent template includes:

[0015] Obtaining a structured data collection template from a database according to the dialogue request;

[0016] Obtaining the user's historical conversation data and the conversation context corresponding to the conversation request;

[0017] Analyzing the historical conversation data and the conversation context using a large model to obtain analysis data;

[0018] The analysis data is used to adjust the questioning method and sequence included in the structured data collection template to obtain a prompt template, and the prompt template is used as a preset intelligent template.

[0019] Optionally, the above method for collecting conversational structured information further includes:

[0020] When receiving an update instruction of the smart template, obtaining an edit field corresponding to the update instruction;

[0021] Analyzing the semantic relationship between the edit field and all fields in the smart template using the large model;

[0022] The smart template is updated according to the semantic relationship and the edit field.

[0023] Optionally, in the above-mentioned method for collecting conversational structured information, determining whether the answer content meets preset answer requirements includes:

[0024] Using a large model to check whether the answer content conforms to a preset format;

[0025] If the answer content conforms to the preset format, it is determined that the answer content meets the preset answer requirements;

[0026] If the answer content does not conform to the preset format, it is determined that the answer content does not meet the preset answer requirements.

[0027] Optionally, the above method for collecting conversational structured information further includes:

[0028] When receiving the repeated collection request sent by the user, determining whether the state of the preset repeated collection switch is on;

[0029] If the state of the preset repeated collection switch is on, returning to execute the dialog request initiated by the receiving user;

[0030] If the state of the preset repeated collection switch is off, a rejection message is fed back to the user.

[0031] Optionally, the above method for collecting conversational structured information further includes:

[0032] If the answer content does not meet the preset answer requirements, or the conversation content contained in the smart template has not been fully collected, then return to executing the business type according to the conversation request and collect the answer content fed back by the user based on the preset smart template.

[0033] A second aspect of the present application provides a conversational structured information collection system, comprising:

[0034] A request receiving unit, configured to receive a conversation request initiated by a user;

[0035] A content collection unit, configured to collect the user's response content based on a preset intelligent template according to the service type of the dialogue request;

[0036] A first judging unit, configured to judge whether the answer content meets a preset answer requirement;

[0037] a second judgment unit, configured to judge whether all conversation contents included in the smart template have been collected if the answer content meets the preset answer requirement;

[0038] an extraction unit configured to obtain all conversation contents if all conversation contents included in the intelligent template have been collected, and extract all answer contents from the conversation contents according to preset rules;

[0039] The conversion unit is used to convert the format of all the answer contents to obtain target answer information, and store the target answer information through the HTTP interface.

[0040] Optionally, the above-mentioned conversational structured information collection system further includes:

[0041] A template acquisition unit, configured to acquire a structured data collection template from a database according to the dialogue request;

[0042] a data acquisition unit, configured to acquire the user's historical conversation data and the conversation context corresponding to the conversation request;

[0043] an analysis unit, configured to analyze the historical conversation data and the conversation context using a large model to obtain analysis data;

[0044] An adjusting unit is used to adjust the questioning method and sequence included in the structured data collection template using the analysis data to obtain a prompt template, and use the prompt template as a preset intelligent template.

[0045] Optionally, the above-mentioned conversational structured information collection system further includes:

[0046] A field acquisition unit, configured to, upon receiving an update instruction of the smart template, acquire an edit field corresponding to the update instruction;

[0047] a relationship analysis unit, configured to analyze the semantic relationship between the edit field and all fields in the smart template using the large model;

[0048] An updating unit is configured to update the smart template according to the semantic relationship and the edit field.

[0049] Optionally, in the above-mentioned conversational structured information collection system, the first judgment unit includes:

[0050] A checking unit, configured to use a large model to check whether the answer content conforms to a preset format;

[0051] a first determining unit, configured to determine whether the answer content meets a preset answer requirement if the answer content conforms to a preset format;

[0052] The second determining unit is configured to determine that the answer content does not meet the preset answer requirement if the answer content does not conform to the preset format.

[0053] Optionally, the above-mentioned conversational structured information collection system further includes:

[0054] a third determining unit, configured to determine whether a preset repeat collection switch is in an on state when receiving a repeat collection request sent by the user;

[0055] a first execution unit, configured to return to executing the dialog request initiated by the receiving user if the preset repeat collection switch is in an on state;

[0056] A feedback unit is configured to feed back rejection information to the user if the preset repeated acquisition switch is in an off state.

[0057] Optionally, the above-mentioned conversational structured information collection system further includes:

[0058] The second execution unit is used to return to execute the business type requested according to the conversation and collect the answer content fed back by the user based on the preset smart template if the answer content does not meet the preset answer requirements or the conversation content contained in the smart template has not been fully collected.

[0059] This application provides a method for collecting conversational structured information. The method receives a conversation request initiated by a user, collects the user's response content based on a preset intelligent template based on the business type of the conversation request, and then determines whether the response content meets the preset response requirements. If the response content meets the preset response requirements, it determines whether all conversation content contained in the intelligent template has been collected. If all conversation content contained in the intelligent template has been collected, it obtains all conversation content and extracts all response content from all conversation content according to preset rules. Finally, it converts all response content into a format to obtain target response information, and stores the target response information through an HTTP interface. Thus, through the flexible guidance of the intelligent agent and the dynamic conversation strategy, users interact naturally with the intelligent assistant during information collection, improving participation and satisfaction, reducing abandonment due to poor experience, and effectively changing the traditional information collection method. BRIEF DESCRIPTION OF THE DRAWINGS

[0060] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are merely embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on the provided drawings without any creative work.

[0061] Figure 1 A flowchart of a method for collecting conversational structured information provided in an embodiment of the present application;

[0062] Figure 2 A schematic diagram of a flow chart of a method for obtaining a preset smart template provided in another embodiment of the present application;

[0063] Figure 3 A flowchart of a method for updating a smart template provided in another embodiment of the present application;

[0064] Figure 4 A flowchart of a method for checking answer content provided in another embodiment of the present application;

[0065] Figure 5 A schematic diagram of the structure of a conversational structured information collection system provided in another embodiment of the present application. DETAILED DESCRIPTION

[0066] The following will be combined with the drawings in the embodiments of this application to clearly and completely describe the technical solutions in the embodiments of this application. Obviously, the embodiments described are only part of the embodiments of this application, not all of the embodiments. Based on the embodiments in this application, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of this application.

[0067] In this application, relational terms such as first and second, etc., are used only to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "comprises," "comprising," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus comprising a series of elements includes not only those elements, but also other elements not explicitly listed, or elements inherent to such process, method, article, or apparatus. In the absence of further limitations, an element defined by the phrase "comprising a ..." does not exclude the presence of additional identical elements in the process, method, article, or apparatus comprising the element.

[0068] An embodiment of the present application provides a method for collecting conversational structured information, which is applied to a system for collecting conversational structured information to solve the problem of poor user experience brought about by the existing technology.

[0069] Specifically, an embodiment of the present application provides a conversational structured information collection system, including: a conversation module, an intelligent agent module, a template setting and management module, an HTTP communication module, a data processing module, and a repeated collection management module.

[0070] Dialogue Module: This module provides a clear and natural dialogue interface for users to interact with the agent. Users can enter responses or request special operations, such as re-collection.

[0071] The intelligent agent module, as the core of the system, asks users questions in an orderly manner based on pre-set intelligent templates and screens their responses. Furthermore, it is responsible for providing clear examples and guidance when users provide unusual responses, ensuring that valid information is obtained. Furthermore, this module determines whether repeated collection operations are permitted based on system settings and takes appropriate action.

[0072] Template Setup and Management: This module allows system administrators to flexibly configure structured field information for collection, adding, modifying, or deleting collection fields at any time to accommodate the data needs of different business scenarios. It also generates corresponding collection templates, laying the foundation for subsequent operations.

[0073] HTTP Communication Module: This module is primarily responsible for communication between the agent and internal systems. The agent uses this module to send HTTP requests to retrieve structured data collection templates. Once the information is collected, the generated JSON data is transferred to the corresponding storage system via HTTP requests, ensuring smooth data transfer and storage.

[0074] Data Processing Module: This module is responsible for extracting, formatting, converting, and storing collected information. It uses a large model to analyze and process conversation content, converting unstructured text information into structured JSON data format according to preset rules to meet the needs of subsequent data utilization and analysis.

[0075] Repeated collection management module: Based on the status of system settings, it determines whether the repeated collection request proposed by the user can be executed, and coordinates the intelligent module and data processing module to complete subsequent operations to ensure the accuracy and integrity of the data.

[0076] Therefore, based on the above-mentioned conversational structured information collection system, the present application embodiment provides a conversational structured information collection method, such as Figure 1 As shown, the specific steps include:

[0077] S101: Receive a conversation request initiated by a user.

[0078] Specifically, the user can open a dialogue window with the intelligent agent of the collection system through the terminal device, and then the user can initiate a dialogue request according to his or her own needs. When the intelligent agent receives the dialogue request initiated by the user, the intelligent agent will automatically identify the user's identity so that subsequent intelligent dialogue can be carried out.

[0079] Optionally, when the agent begins working, it must first deploy the appropriate computing and storage resources on the server to meet the large model's operational and data storage requirements. Furthermore, ensure that the server and terminal devices (such as the user's computer or mobile phone) are properly connected via the network. Next, install and configure the software libraries and frameworks related to the large model, as well as the development and runtime environment for each system module involved in this application. Initialize the agent, including loading pre-trained large model parameters and defining an initial acquisition template.

[0080] S102: According to the service type of the dialogue request, the user's response content based on the preset intelligent template is collected.

[0081] It should be noted that in this embodiment of the application, the dialogue module not only supports text and voice input, but also further integrates multimodal information standby methods such as image recognition. Therefore, in order to improve the comprehensiveness and accuracy of the collected information, the intelligent agent in this embodiment of the application can first obtain the corresponding preset intelligent template based on the business type of the dialogue request. The intelligent agent then asks questions to the user in sequence based on the questions in the intelligent template. For example, "Hello, please tell me your name." The user can then type or speak to answer the question in the dialogue window provided by the collection system, so that the intelligent agent can collect the user's response based on the question feedback.

[0082] It's also worth noting that the constructed knowledge graph can also be utilized. That is, when the agent is in conversation with the user, it can use the information it has already collected to expand the associations in the knowledge graph and proactively inquire about important information related to it that the user has not yet mentioned. For example, in a business information collection scenario, after collecting the business scope of the enterprise, the agent can further collect information based on common business models and industry trends related to that business scope in the knowledge graph, thereby increasing the depth and breadth of the acquired data.

[0083] Optionally, in another embodiment of the present application, a method for obtaining a preset smart template is provided, such as Figure 2 As shown, the specific steps include:

[0084] S201. Obtain a structured data collection template from a database according to a dialogue request.

[0085] Specifically, the intelligent agent will first determine the type of template that needs to be obtained based on the business type of the dialogue request, then build an HTTP request based on the dialogue request, and then the intelligent agent sends the HTTP request to the designated internal system through the HTTP communication module. Finally, the internal system processes the HTTP request and feeds back the corresponding structured data collection template from the database to the intelligent agent, usually in JSON format.

[0086] Optionally, the agent can store the acquired structured data collection template in local memory or database to prepare for subsequent data collection tasks.

[0087] S202: Obtain the user's historical conversation data and the conversation context corresponding to the conversation request.

[0088] It should be noted that in order for the intelligent agent to better understand the user's preferences, needs and tone, and thus provide more customized responses and improve the user experience, it can be analyzed by obtaining the user's historical conversation data. In order for the intelligent agent to more accurately understand the user's intentions and provide answers that better meet the user's needs, it can also obtain the conversation context corresponding to the conversation request. Therefore, by pre-acquiring historical conversation data and conversation context, the intelligent agent can become more intelligent and efficient in its interaction with the user.

[0089] S203: Analyze historical conversation data and conversation context using the big model to obtain analysis data.

[0090] It is understandable that the large model has deep semantic understanding capabilities, so it can analyze the user's historical conversation data (if any) and the current conversation context, so that the way and order of asking questions can be dynamically adjusted later.

[0091] Specifically, by using big models to analyze historical data and conversation context, users' common request patterns and topic preferences can be identified. The intelligent agent can prepare relevant information in advance, thereby reducing unnecessary repeated questions and conversation waiting time, improving conversation efficiency, and through the contextual analysis of historical conversations by big models, the big model can learn from users' emotional reactions, expressions and conversation expectations in specific scenarios, thereby optimizing conversation quality and ensuring that responses are more targeted and accurate. Finally, through in-depth analysis of historical conversation data, the big model can predict users' possible subsequent needs or interests, prepare and recommend related content or services in advance, enhance user stickiness and satisfaction, and integrate the results of the analysis to obtain analysis data.

[0092] For example, when collecting medical information, if the agent discovers that the user's description of symptoms is vague and that the user has difficulty understanding professional terminology in previous conversations, the agent will not only provide examples to guide the user but also adjust subsequent questions to ask for relevant information in more accessible terms, ensuring that the user can accurately understand and answer. This dynamic adjustment strategy is based on the innovative application of the contextual understanding capabilities of large models, effectively improving the efficiency and accuracy of information collection. Existing technologies lack this flexible adjustment mechanism based on user characteristics and conversational context.

[0093] S204: Using the analysis data, adjust the questioning method and sequence included in the structured data collection template to obtain a prompt template, and use the prompt template as a preset intelligent template.

[0094] Specifically, through the analysis of historical data and conversational context, we found which questioning methods (such as open questions, closed questions, etc.) and sequences can best stimulate user interest and participation, and identified which questions are likely to arouse positive or negative emotions in users, thereby adjusting the tone and form of questions in the structured data collection template to improve user participation and satisfaction.

[0095] Furthermore, you can optimize the order of questions from simple to complex: Based on the user's receptiveness to questions, gradually guide the questions from basic and simple content to complex and in-depth content to avoid user rejection or fatigue. And optimize the questioning method, that is, use appropriate question types: Based on the different needs of users, choose the appropriate question type (for example: multiple-choice questions, short-answer questions, rated questions, etc.) to improve the accuracy and enthusiasm of users' answers.

[0096] Finally, based on the above analysis results, a specific prompt template is designed to ensure that the template's questioning method, sequence, and content are in line with the optimal design, so that the prompt template can be adaptively adjusted according to different situations and user feedback, used to drive multiple rounds of dialogue with users, and realize information collection, extraction, and storage operations. The optimized prompt template is saved as a standard template as the basic framework for data collection.

[0097] Optionally, the intelligent agent module can also automatically detect updates to the structured data collection template and regenerate the prompt template (intelligent template) based on the new template. In subsequent information collection dialogues, it will automatically ask users questions related to their allergy history, thereby achieving dynamic expansion of collection fields without the need for large-scale modifications to the entire system to meet new business development needs.

[0098] Optionally, after obtaining the smart template, the collection system allows the system administrator to flexibly configure the structured field information in the smart template, and supports adding, modifying or deleting collection fields at any time according to different business needs, thereby realizing dynamic update of the smart template. Therefore, in another embodiment of the present application, a method for updating the smart template is also provided, such as Figure 3 As shown, the specific steps include:

[0099] S301: When an update instruction of a smart template is received, an edit field corresponding to the update instruction is obtained.

[0100] The update instruction may include an add instruction, a modify instruction, or a delete instruction. It should be noted that each instruction corresponds to a modification of a collection field. Therefore, in order to flexibly and effectively modify the collection fields in the smart template later, it is necessary to obtain the edit field corresponding to the update instruction in advance.

[0101] S302: Analyze the semantic relationship between the edit field and all fields in the smart template using the big model.

[0102] Specifically, large models (such as BERT and GPT) are used to perform semantic embedding on edit fields and all fields in smart templates, converting each field into a vector representation that captures the semantic information of the field content. Secondly, the constructed knowledge graph is used to identify the relationships between fields. This graph captures the semantic connections between fields by analyzing the word meaning, context, and grammatical structure of the fields. Natural language processing techniques (such as dependency parsing and entity relationship extraction) are then used to extract the potential semantic relationships between edit fields and all fields in smart templates. These relationships may include field dependencies (such as the association between order numbers and customer names) or interactions between fields (such as the time series relationship between date fields and status fields).

[0103] Optionally, semantically similar fields can be identified by calculating the similarity between field vectors. For example, user names and contact names may be highly semantically similar, and association rule mining techniques can be used to analyze potential associations between fields.

[0104] S303: Update the smart template according to the semantic relationship and edit fields.

[0105] Specifically, the display order of fields in the smart template is adjusted based on the dependencies between them. For example, the order number is displayed first, followed by the customer name, as the customer name may be automatically populated based on the order number. Finally, the system can automatically generate certain fields based on the identified semantic relationships.

[0106] For example, taking the collection of e-commerce user information as an example, when a new field called "frequently purchased categories" is added, the system finds through large-scale model analysis that this field is closely related to fields such as "purchase frequency" and "consumption preferences". Therefore, when generating the prompt template, these related questions are reasonably associated to make the dialogue process more logical and coherent. This intelligent association expansion function solves the problem of logical integration when traditional form-based collection methods are difficult to cope with field changes, and is an innovation of the present invention in collection template management. For example, if a user mentions "diabetes", the semantic network automatically associates fields such as "blood sugar value", "medication history" and "complications". The intelligent agent can dynamically add questions without manual intervention to form a semantic chain of "main question-related question".

[0107] S103: Determine whether the answer content meets the preset answer requirements.

[0108] It is understandable that in order to promptly correct information that does not meet the requirements and help users express themselves accurately through guided examples, thereby ensuring that the collected data is complete and accurate, reducing the workload of subsequent data cleaning and correction, and improving data availability, it is necessary to determine whether the answer content of the user feedback meets the preset answer requirements. If the answer content meets the preset answer requirements, execute step S104.

[0109] Optionally, after executing step S103, the method further includes:

[0110] If the answer content does not meet the preset answer requirements, it returns to execute the business type according to the dialogue request and collects the answer content fed back by the user based on the preset intelligent template.

[0111] It is understandable that when the answer content does not meet the preset answer requirements, for example, the user enters "12345" as the name, the intelligent agent will provide the user with real-time examples (such as "Please enter your real name, such as 'Zhang San'") according to the preset rules, guide the user to answer again, and improve the guidance effect until valid information is obtained, that is, return to execute step S102.

[0112] Optionally, in another embodiment of the present application, a specific implementation of step S103 is as follows: Figure 4 As shown, the specific steps include:

[0113] S401. Use the big model to check whether the answer content conforms to the preset format.

[0114] It should be noted that in order to ensure the integrity and accuracy of the data and reduce subsequent cleaning and correction work, the system should promptly correct information that does not meet the requirements and help users express themselves accurately through guided examples, thereby improving the usability of the data. Therefore, a large model will be used to check whether the answer content meets the expected format and requirements. That is, the large model is used to perform grammatical analysis and semantic understanding on the content input by the user. The input content is first broken down into specific fields, phrases, sentences, etc., and compared with the preset format. Then, according to the preset format, it is checked whether the fields in the input content appear in the correct order. Finally, it is checked whether the type of each field meets the expectations, such as numbers, dates, text, etc. Therefore, if the answer content meets the preset format, step S402 is executed. If the answer content does not meet the preset format, step S403 is executed.

[0115] S402: Determine whether the answer content meets the preset answer requirements.

[0116] S403: Determine whether the answer content does not meet the preset answer requirements.

[0117] Specifically, when the answer content does not conform to the preset format, the intelligent agent needs to guide the user to answer again until valid information is obtained.

[0118] S104: Determine whether all conversation contents included in the intelligent template have been collected.

[0119] It should be noted that to ensure complete collection of conversation content and provide accurate and reliable basic data for subsequent data processing, analysis, and decision-making, and to ensure that all necessary responses are collected, it is necessary to determine whether all conversation content included in the smart template has been fully collected. If it is determined that all conversation content included in the smart template has been fully collected, step S105 is executed.

[0120] Optionally, after executing step S104, the method further includes:

[0121] If the conversation content included in the smart template has not been fully collected, the process returns to execute the service type according to the conversation request and collects the user's response content based on the preset smart template.

[0122] It should be noted that after obtaining the answer content that meets the preset requirements, the intelligent agent continues to ask the next question in the order of the collection template, such as "How old are you?" Repeat the above answering, judgment, and exception handling process, gradually collect complete user information, and then return to execute step S102.

[0123] In addition, when all questions are answered effectively, the agent determines that the information collection process is completed.

[0124] S105: Obtain all conversation contents, and extract all answer contents from all conversation contents according to preset rules.

[0125] It should be noted that once all conversations included in the smart template have been collected, the data processing module needs to extract the user's responses according to preset rules for subsequent conversion into formatted JSON data. This involves first segmenting all conversations into distinct stages or sentences for subsequent processing. Then, based on requirements, preset rules for extracting responses are defined. For example, the following rules may be used:

[0126] The format of the answer: whether it is text, number, multiple-choice list, etc.

[0127] Position of the answer: Does the answer always immediately follow the question, or is there some other marker?

[0128] Specific identification: Are there specific keywords marking the answer section (for example, "Answer:", "Result:", etc.).

[0129] Rule Type:

[0130] Keyword matching: Extract relevant answers by searching for specific keywords.

[0131] Grammatical analysis: Analyze sentence structure and extract the core information in the sentence.

[0132] Contextual understanding: Determine which responses are valid based on the conversation context.

[0133] Then, according to the defined preset rules, all conversation contents are checked one by one, and the answers that meet the rules are extracted.

[0134] If the rule is based on keywords, the answer can be extracted by searching for specific keywords in the conversation content.

[0135] If the rule is context-based, it analyzes the user's answer and the agent's question to determine whether the user's response is an answer.

[0136] Finally, these rules are automatically applied through the data processing module to extract the effective answers in each conversation.

[0137] S106: Convert all answer contents into different formats to obtain target answer information, and store the target answer information through an HTTP interface.

[0138] Understandably, to facilitate subsequent data storage, querying, analysis, and mining, and to better leverage the value of the collected data, providing a strong basis for enterprise decision support and business optimization, all responses need to be converted into formatted JSON data. This data format is well-structured and facilitates subsequent data processing. The resulting JSON data is then sent via an HTTP communication module to the business system database corresponding to the business type. The data is then saved to the corresponding business system database, such as a hospital's information management system, allowing subsequent medical staff to access and utilize the data for diagnosis, treatment, and other tasks.

[0139] For example, the following JSON data is generated:

[0140] {

[0141] "name":"Zhang San",

[0142] "age":25,

[0143] "gender":"male",

[0144] "symptoms": "headache and fever for two days",

[0145] }.

[0146] The field names of the JSON data (such as "name", "age", "gender", etc.) are derived from the definition of the collection template and are clearly reflected in the automatically generated prompt template.

[0147] Optionally, the user can re-initiate a conversation request according to their own needs. Therefore, in order to promptly respond to the user's repeated conversation request and avoid affecting the user's experience, another embodiment of the present application further provides a method for repeated collection, which specifically includes the following steps:

[0148] When a repeated collection request sent by a user is received, it is determined whether the state of the preset repeated collection switch is on.

[0149] It should be noted that when a user types or speaks "I want to re-enter the information" in the collection system's dialog window, the agent receives the user's request for a repeat collection through the dialog window interface. The agent then checks the system's repeat collection switch status. This is because properly managing repeat collection operations not only meets the user's need to re-collect when necessary to correct errors or update information, but also avoids data confusion and resource waste caused by frequent repeated collections, ensuring data timeliness and accuracy. Therefore, the agent then determines whether the preset repeat collection switch is on. If it is on, indicating that repeated collection is allowed, the agent returns to step S102, restarting the information collection process, questioning the user again, and overwriting the existing record with the new data after the collection is completed. If the preset repeat collection switch is off, indicating that repeated collection is not allowed, the agent will then provide the user with a rejection message. This rejection message may be something like, "Sorry, this information collection has been completed. Repeated collection is not allowed. Please ensure that the information you provided is accurate. If you have any questions, please consult the relevant staff." This politely rejects the user's request and enhances the user experience.

[0150] This application provides a method for collecting conversational structured information. The method receives a conversation request initiated by a user, collects the user's response content based on a preset intelligent template based on the business type of the conversation request, and then determines whether the response content meets the preset response requirements. If the response content meets the preset response requirements, it determines whether all conversation content contained in the intelligent template has been collected. If all conversation content contained in the intelligent template has been collected, it obtains all conversation content and extracts all response content from all conversation content according to preset rules. Finally, it converts all response content into a format to obtain target response information, and stores the target response information through an HTTP interface. Thus, through the flexible guidance of the intelligent agent and the dynamic conversation strategy, users interact naturally with the intelligent assistant during information collection, improving participation and satisfaction, reducing abandonment due to poor experience, and effectively changing the traditional information collection method.

[0151] Another embodiment of the present application provides a conversational structured information collection system, such as Figure 5 As shown, it includes the following units:

[0152] The request receiving unit 501 is configured to receive a conversation request initiated by a user.

[0153] The content collection unit 502 is used to collect the answer content fed back by the user based on the preset intelligent template according to the service type of the dialogue request.

[0154] The first judging unit 503 is configured to judge whether the answer content meets the preset answer requirements.

[0155] The second judgment unit 504 is configured to judge whether all the conversation contents included in the intelligent template have been collected if the answer content meets the preset answer requirement.

[0156] The extraction unit 505 is configured to obtain all the conversation contents if all the conversation contents included in the intelligent template have been collected, and extract all the answer contents from all the conversation contents according to a preset rule.

[0157] The conversion unit 506 is used to convert the format of all answer contents to obtain target answer information, and store the target answer information through the HTTP interface.

[0158] It should be noted that the specific working process of the above modules in the embodiment of the present application can refer to steps S101 to S106 in the above method embodiment, and will not be repeated here.

[0159] Optionally, another embodiment of the present application provides a system for collecting conversational structured information, further comprising:

[0160] The template acquisition unit is used to acquire the structured data collection template from the database according to the dialogue request.

[0161] The data acquisition unit is used to obtain the user's historical conversation data and the conversation context corresponding to the conversation request.

[0162] The analysis unit is used to analyze historical conversation data and conversation context using a large model to obtain analysis data.

[0163] The adjustment unit is used to adjust the questioning method and sequence contained in the structured data collection template by using the analysis data to obtain a prompt template, and use the prompt template as a preset intelligent template.

[0164] Optionally, another embodiment of the present application provides a system for collecting conversational structured information, further comprising:

[0165] The field acquisition unit is used to acquire the edit field corresponding to the update instruction when receiving the update instruction of the smart template.

[0166] The relationship analysis unit is used to analyze the semantic relationship between the edit field and all fields in the smart template using the big model.

[0167] The update unit is used to update the smart template based on semantic relationships and edit fields.

[0168] Optionally, in a conversational structured information collection system provided by another embodiment of the present application, the first judgment unit 503 includes:

[0169] The checking unit is used to use the large model to check whether the answer content conforms to the preset format.

[0170] The first determining unit is configured to determine that the answer content meets a preset answer requirement if the answer content conforms to a preset format.

[0171] The second determining unit is configured to determine that the answer content does not meet the preset answer requirement if the answer content does not conform to the preset format.

[0172] Optionally, another embodiment of the present application provides a system for collecting conversational structured information, further comprising:

[0173] The third judgment unit is configured to judge whether the state of the preset repeated collection switch is on when receiving the repeated collection request sent by the user.

[0174] The first execution unit is configured to return to execute the dialog request initiated by the receiving user if the preset repeat collection switch is in the on state.

[0175] The feedback unit is used to feed back rejection information to the user if the state of the preset repeated collection switch is off.

[0176] Optionally, another embodiment of the present application provides a system for collecting conversational structured information, further comprising:

[0177] The second execution unit is used to return to execute the business type according to the conversation request and collect the answer content fed back by the user based on the preset smart template if the answer content does not meet the preset answer requirements or the conversation content contained in the smart template has not been fully collected.

[0178] It should be noted that the specific working processes of the various modules provided in the above embodiments of the present application can refer to the corresponding steps in the above method embodiments, and will not be repeated here.

[0179] It should also be noted that the embodiment of the present application provides a conversational structured information collection system, which has the technical effects of any of the above embodiments, and the embodiment of the present application will not be described in detail here.

[0180] Professionals may further appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of the two. In order to clearly illustrate the interchangeability of hardware and software, the above description has generally described the components and steps of each example according to their functions. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professionals and technicians may use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0181] The above description of the disclosed embodiments is intended to enable one skilled in the art to implement or use the present application. Various modifications to these embodiments will be readily apparent to one skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present application. Therefore, the present application is not limited to the embodiments shown herein, but is intended to conform to the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A method for collecting conversational structured information, characterized in that: include: Receive conversation requests initiated by users; According to the business type of the dialogue request, collecting the user's response content based on the preset intelligent template; Determine whether the answer content meets the preset answer requirements; If the answer content meets the preset answer requirements, then determine whether all the conversation content included in the smart template has been collected; If all the conversation contents included in the smart template have been collected, all the conversation contents are obtained, and all the answer contents are extracted from the conversation contents according to the preset rules; All the answer contents are formatted to obtain target answer information, and the target answer information is stored through an HTTP interface.

2. The method according to claim 1, characterized in that The method for obtaining the preset intelligent template includes: Obtaining a structured data collection template from a database according to the dialogue request; Obtaining the user's historical conversation data and the conversation context corresponding to the conversation request; Analyzing the historical conversation data and the conversation context using a large model to obtain analysis data; The analysis data is used to adjust the questioning method and sequence included in the structured data collection template to obtain a prompt template, and the prompt template is used as a preset intelligent template.

3. The method according to claim 2, characterized in that Also includes: When receiving an update instruction of the smart template, obtaining an edit field corresponding to the update instruction; Analyzing the semantic relationship between the edit field and all fields in the smart template using the large model; The smart template is updated according to the semantic relationship and the edit field.

4. The method according to claim 1, wherein The determining whether the answer content meets the preset answer requirements includes: Using a large model to check whether the answer content conforms to a preset format; If the answer content conforms to the preset format, it is determined that the answer content meets the preset answer requirements; If the answer content does not conform to the preset format, it is determined that the answer content does not meet the preset answer requirements.

5. The method according to claim 1, characterized in that Also includes: When receiving the repeated collection request sent by the user, determining whether the state of the preset repeated collection switch is on; If the state of the preset repeated collection switch is on, returning to execute the dialog request initiated by the receiving user; If the state of the preset repeated collection switch is off, a rejection message is fed back to the user.

6. The method according to claim 1, characterized in that Also includes: If the answer content does not meet the preset answer requirements, or the conversation content contained in the smart template has not been fully collected, then return to executing the business type according to the conversation request and collect the answer content fed back by the user based on the preset smart template.

7. A conversational structured information collection system, characterized in that: include: A request receiving unit, configured to receive a conversation request initiated by a user; A content collection unit, configured to collect the user's response content based on a preset intelligent template according to the service type of the dialogue request; A first judging unit, configured to judge whether the answer content meets a preset answer requirement; a second judgment unit, configured to judge whether all conversation contents included in the smart template have been collected if the answer content meets the preset answer requirement; an extraction unit configured to obtain all conversation contents if all conversation contents included in the intelligent template have been collected, and extract all answer contents from the conversation contents according to preset rules; The conversion unit is used to convert the format of all the answer contents to obtain target answer information, and store the target answer information through the HTTP interface.

8. The system according to claim 7, characterized in that Also includes: A template acquisition unit, configured to acquire a structured data collection template from a database according to the dialogue request; a data acquisition unit, configured to acquire the user's historical conversation data and the conversation context corresponding to the conversation request; an analysis unit, configured to analyze the historical conversation data and the conversation context using a large model to obtain analysis data; An adjusting unit is used to adjust the questioning method and sequence included in the structured data collection template using the analysis data to obtain a prompt template, and use the prompt template as a preset intelligent template.

9. The system according to claim 8, characterized in that Also includes: A field acquisition unit, configured to, upon receiving an update instruction of the smart template, acquire an edit field corresponding to the update instruction; a relationship analysis unit, configured to analyze the semantic relationship between the edit field and all fields in the smart template using the large model; An updating unit is configured to update the smart template according to the semantic relationship and the edit field.

10. The system according to claim 7, wherein: The first judgment unit includes: A checking unit, configured to use a large model to check whether the answer content conforms to a preset format; A first determining unit, configured to determine whether the answer content meets a preset answer requirement if the answer content conforms to a preset format; The second determining unit is configured to determine that the answer content does not meet the preset answer requirement if the answer content does not conform to the preset format.

Images