A personal number knowledge base creation and configuration method supporting multi-dimensional screening

Abstract

The application discloses a kind of personal number knowledge base creation and configuration method supporting multi-dimension screening, it is related to data processing technical field, method includes: creating and the exclusive knowledge base of personal number binding, establish and the data interface connection of customer management system, communication push system;Configuration knowledge base basic function, establish file and the initial mapping of customer label, business scenario;Multi-dimension screening logic containing basic attribute, customer behavior, cross-system feedback dimension is built, preset screening dimension self-evolution process, business conflict calibration process and cross-system screening dimension generation process are handled;The process is executed to realize the real-time linkage of screening result and customer communication scene, multi-system, complete predictive screening and business process dynamic calibration.This method improves the adaptability and efficiency of knowledge resource screening, and provides support for personal number to efficiently carry out business.

Description

Technical Field

[0001] This invention relates to the field of data processing technology, and in particular to a method for creating and configuring a personal account knowledge base that supports multi-dimensional filtering. Background Technology

[0002] With the widespread use of personal accounts in business communication, customer service and other scenarios, the knowledge base accompanying personal accounts has become the core data carrier supporting business operations. It needs to meet the needs of personal account users to quickly access and adapt knowledge resources in customer communication and multi-system integration.

[0003] The existing filtering logic of personal account knowledge bases is mostly based on basic file attributes, resulting in a relatively singular filtering dimension. Furthermore, the filtering function lacks deep integration with the personal account's customer communication scenarios and its connection to multiple business systems. It can only achieve basic file location and cannot dynamically adjust filtering strategies based on customer behavior characteristics or cross-system business feedback. Simultaneously, the existing filtering logic lacks predictive filtering capabilities and cannot dynamically calibrate deviations in personal account business processes. This leads to insufficient adaptability of filtering results to actual business needs, low efficiency in knowledge resource retrieval, and difficulty in supporting personal accounts to efficiently conduct customer communication and multi-system collaborative business, thus hindering business expansion and service quality improvement for personal accounts. Summary of the Invention

[0004] To address the technical problems existing in the prior art, this invention provides a method for creating and configuring a personal account knowledge base that supports multi-dimensional filtering.

[0005] The technical solution adopted in this invention is: a method for creating and configuring a personal account knowledge base that supports multi-dimensional filtering, comprising the following steps:

[0006] Step 1: Receive the user's input instruction to associate a personal account, create a dedicated knowledge base bound to the personal account, and simultaneously establish communication connections between the personal account and the customer management data interface and the multi-business system data interface. The multi-business system includes at least the customer management system and the communication push system.

[0007] Step 2: Receive user input of basic knowledge base configuration instructions, and execute the logic for uploading configuration files, displaying file lists, deleting files, and importing multi-format files. The file upload logic supports batch upload and single upload, and the multi-format files must cover at least documents, audio, video, and images. At the same time, an initial mapping relationship between files and customer tags and business scenarios is established through data association algorithms.

[0008] Step 3: Receive the user's input instruction to build an intelligent filtering system, and build a multi-dimensional filtering logic that includes basic attribute dimensions, customer behavior dimensions, and cross-system feedback dimensions. The multi-dimensional filtering logic includes a preset filtering dimension self-evolution processing flow, a business conflict calibration processing flow, and a cross-system filtering dimension generation processing flow.

[0009] Step 4: Receive the user's input filtering business linkage instruction, and sequentially execute the filtering dimension self-evolution processing flow, business conflict calibration processing flow, and cross-system filtering dimension generation processing flow to realize real-time data linkage between filtering results and personal account customer communication scenarios and multiple business systems, and complete the predictive generation of filtering results and dynamic calibration of personal account business processes.

[0010] Preferably, the self-evolutionary processing flow for the filtering dimensions includes the following:

[0011] Receive instructions to collect customer communication behavior data from personal accounts, and obtain customer communication behavior data such as call frequency, consultation keywords, and interaction scenario preferences through real-time data capture, and store it in the behavior database;

[0012] The dimension priority algorithm is invoked to automatically adjust the priority order of each filtering dimension based on the customer communication behavior data in the behavior database; when high-frequency call customer data is detected, the high-value tags in the customer tags are automatically set as the default filtering conditions, and the file data containing the consultation keywords are extracted first through the keyword matching algorithm.

[0013] Receive instructions to generate predicted filtering results, call the historical communication data analysis model, predict customer needs based on historical communication data, automatically filter matching file data according to the prediction results, and display the filtered results in a floating format on the personal account interface.

[0014] Preferably, the business conflict calibration process includes the following:

[0015] Receive a business scenario conflict identification instruction, call the conflict analysis algorithm, and identify the conflict type by parsing the filtering results, using feedback data and business system format compatibility data. The conflict type includes at least content adaptation conflict and format compatibility conflict.

[0016] Upon receiving the filtering logic calibration command, for content adaptation conflicts, it automatically increases the weight parameters of the corresponding filtering dimensions for video files and simultaneously generates optimized prompt data for file format conversion; for format compatibility conflicts, it calls the file content matching algorithm to filter alternative file data that is consistent with the original file content and adds format incompatibility identification information to the original file data.

[0017] The system receives instructions to generate business process optimization suggestions, and generates personal account business process optimization suggestion data based on conflict record data in the historical conflict database. After receiving instructions from the user to confirm the adoption of the optimization suggestions, the system automatically updates the format compatibility rule parameters in the filtering logic.

[0018] Preferably, the cross-system filtering dimension generation process includes the following:

[0019] Receive feedback collection instructions from multiple business systems, establish data interaction with the customer management system and communication push system respectively, obtain customer conversion rate data after file viewing and SMS open rate data corresponding to the file, and store them in the cross-system feedback database;

[0020] Upon receiving the instruction to adjust the weight of the filtering dimension, the threshold comparison algorithm is invoked. If the customer conversion rate data corresponding to the file reaches the preset high conversion threshold, the weight value of the knowledge type dimension of the file is automatically increased; if the SMS open rate data corresponding to the file is lower than the preset effective open rate threshold, the priority parameter of the file in the idle SMS push business scenario is automatically reduced.

[0021] The system receives a cross-system filtering dimension generation instruction, calls a dimension combination algorithm, and generates new filtering dimensions based on multi-business system feedback data in the cross-system feedback database. The new filtering dimensions include at least a combination of dimensions representing high conversion and high open rate. After receiving the filtering instruction input by the user, the system completes the filtering of file data based on the new filtering dimensions.

[0022] Preferably, the file upload includes the following:

[0023] The system receives file behavior association instructions from the user, calls the historical data association algorithm, automatically associates the historical customer communication behavior data of the personal account when uploading file data, and initially assigns corresponding customer tags to the file data through tag matching.

[0024] The system receives an instruction to automatically generate an index, and generates cross-system call index data based on customer tags and business scenario association attributes of file data. The cross-system call index data is used to quickly locate file data in the subsequent filtering process.

[0025] Preferably, the multi-dimensional filtering logic further includes a filtering rule template configuration process, which includes the following:

[0026] Receive the user's input instruction to save the filter template, and store the combined logic, which includes filter conditions based on customer behavior and cross-system feedback, in the template database as a template.

[0027] Receive template self-optimization instructions, monitor customer behavior changes and cross-system feedback updates in real time, and call template adjustment algorithms to automatically update the filter condition parameters in the filter template.

[0028] Preferably, the multi-dimensional filtering logic further includes a filtering permission hierarchical control process, which includes the following:

[0029] Receive user input of role behavior permission binding instructions, assign corresponding operation permission data according to personal account roles, the personal account roles include at least creator, administrator and ordinary user, and establish a mapping relationship between each role and customer communication behavior data viewing permissions;

[0030] Upon receiving a permission conflict warning command, the system invokes a permission verification algorithm. If it detects that a role has initiated a request to filter highly sensitive files that exceed its permission scope, it automatically triggers a warning signal and intercepts the filtering request. At the same time, it generates permission request prompt data and pushes it to the user interface.

[0031] Preferably, the real-time data linkage between the filtering results and customer communication scenarios via personal accounts, as well as across multiple business systems, includes the following:

[0032] Receive batch association instructions from users, call the scenario matching algorithm, and match and recommend suitable business scenario data based on cross-system feedback data to the filtering results;

[0033] Receive batch activation status configuration instructions, provide activation method options for the filtered results, the activation methods include at least immediate activation and activation for a specified time period; call the time period matching algorithm to automatically verify the overlap between the user's selected specified time period and the customer's active peak time period determined based on historical call data, and output the verification results to help improve the file reach rate.

[0034] Preferably, the dynamic calibration of the personal account business process includes the following:

[0035] Upon receiving a business process calibration trigger instruction, the system identifies the deviation types in the personal account business process by parsing the linkage data between the filtering results and the business scenario. The deviation types include at least filtering result delivery delay deviation and business scenario adaptation deviation. The system then invokes a process optimization algorithm to generate a business process calibration plan based on the deviation types and historical linkage data. The calibration plan includes at least filtering result push timing adjustment parameters and business scenario matching rule update suggestions. After receiving a user confirmation instruction for calibration, the system automatically executes the calibration plan and synchronously stores the process parameters and business effect data before and after calibration in the process calibration database for subsequent optimization and iteration of the calibration plan.

[0036] Preferably, the process of generating personal account business process optimization suggestion data based on conflict record data in the historical conflict database further includes the following:

[0037] The conflict record data in the historical conflict database is classified and organized, and a classification index is established according to the conflict type, which includes content adaptation conflict and format compatibility conflict.

[0038] The calibration schemes corresponding to various conflicts, the calibrated business feedback data, and the conflict type classification index are associated and stored to form an association record. When the same type of business scenario conflict is detected again, the corresponding association record is quickly retrieved through the classification index, and new business process optimization suggestion data is generated based on the historical calibration schemes and business feedback data.

[0039] The beneficial effects of the present invention are at least one of the following:

[0040] By constructing a multi-dimensional filtering logic that includes basic attribute dimensions, customer behavior dimensions, and cross-system feedback dimensions, compared with the traditional single-dimensional filtering method, the filtering dimensions of the personal account knowledge base are effectively enriched, and the adaptability of the filtering results to the actual business needs of the personal account is improved.

[0041] By establishing communication connections between personal accounts and customer management data interfaces, as well as data interfaces of multiple business systems, and linking the filtering results with real-time data from personal account customer communication scenarios and multiple business systems, the synergy between the filtering function and business scenarios and related systems is strengthened, which helps to improve the targeting of knowledge resource retrieval.

[0042] By leveraging a self-evolving processing flow for filtering dimensions to achieve predictive generation of filtering results, users can quickly retrieve suitable knowledge resources in customer communication scenarios, reducing resource search costs.

[0043] By calibrating the business conflict handling process and dynamically calibrating the personal account business process, the matching degree between the screening logic and the business process can be optimized in a timely manner, which helps to improve the smoothness and efficiency of personal account business operations. Attached Figure Description

[0044] Figure 1 This is a flowchart illustrating the method of the present invention. Detailed Implementation

[0045] The embodiments of the present invention will now be described in detail with reference to the accompanying drawings.

[0046] Example 1

[0047] Existing personal account knowledge bases only support basic file attribute filtering, lacking specific dimensions such as customer behavior and cross-system feedback. Furthermore, the filtering results cannot be linked in real-time with customer communication scenarios or related business systems, resulting in insufficient adaptability and low efficiency in knowledge resource retrieval, making it difficult to support efficient business operations for personal accounts. To address this technical problem, this embodiment provides a method for creating and configuring a personal account knowledge base that supports multi-dimensional filtering, such as... Figure 1 As shown, it includes the following steps:

[0048] Step 1: Receive the user's input instruction to associate a personal account, create a dedicated knowledge base bound to the personal account, and simultaneously establish communication connections between the personal account and the customer management data interface and the multi-business system data interface. The multi-business system includes at least the customer management system and the communication push system.

[0049] For example, the computer receives a personal account association instruction entered by the user through the personal account management backend, such as input by the account manager. This instruction includes the unique identifier of the personal account, such as the exclusive service number code, and information such as the associated business type, such as enterprise product sales services.

[0050] Based on the above instructions, the computer creates a unique knowledge base bound to the personal account. The knowledge base is only visible to the personal account and authorized users, ensuring the exclusive management of knowledge resources.

[0051] Simultaneously establish communication connections between the aforementioned personal accounts and the data interfaces of the customer management system and the communication push system, using standard API interface protocols to achieve data interaction. Specifically, the interface with the customer management system is used to obtain customer tags, such as the customer's industry, cooperation level, and historical communication records; the interface with the communication push system is used to implement the linkage of SMS or flash message pushes of filtering results, such as product information and coupon information.

[0052] Step 2: Receive user input of basic knowledge base configuration commands, including configuration file upload, file list display, file deletion, and multi-format file import. The file upload logic supports batch upload and single upload, and the multi-format files must cover at least documents, audio, video, and images. At the same time, an initial mapping relationship between files and customer tags and business scenarios is established through a data association algorithm.

[0053] For example, the computer receives basic knowledge base configuration instructions from the user and configures the execution logic of core basic functions according to the instructions:

[0054] Configuration file upload logic: Supports drag-and-drop upload of single files and batch upload of multiple files. The upload interface is compatible with various file formats, including document formats such as doc and pdf, audio formats such as mp3 and wav, video formats such as mp4 and avi, and image formats such as jpg and png.

[0055] Configuration file list display and deletion logic: The file list is displayed in reverse order of upload time by default, supports fuzzy search by file name, deletion operation requires secondary confirmation, and deleted data is retained for a 30-day backup period;

[0056] Configure multi-format file compatibility import logic: Use a format parsing plugin to uniformly encode different types of files, ensuring that core file attributes, such as document keywords and video duration, can be extracted during subsequent filtering;

[0057] The computer invokes a data association algorithm to establish an initial mapping relationship between files, customer tags, and business scenarios:

[0058] Customer tag mapping: Based on customer classification data synchronized from the customer management system, preset tag items such as industry type, cooperation level, and demand scenario are available. When users upload files, they can manually select associated tags, or the system can automatically match and recommend tags based on the file content. For example, when uploading a chemical industry product manual, the system recommends chemical industry tags.

[0059] Business scenario mapping: Preset options for business scenarios such as customer inquiry response, cooperation negotiation support, after-sales follow-up, and SMS push notifications. Users must specify at least one applicable scenario when uploading files, laying the foundation for subsequent scenario-based filtering.

[0060] Step 3: Receive the user's input instruction to build an intelligent filtering system, and build a multi-dimensional filtering logic that includes basic attribute dimensions, customer behavior dimensions, and cross-system feedback dimensions. The multi-dimensional filtering logic includes a preset filtering dimension self-evolution processing flow, a business conflict calibration processing flow, and a cross-system filtering dimension generation processing flow.

[0061] For example, the computer receives user input instructions for building an intelligent filtering system, and constructs multi-dimensional filtering logic that includes basic attribute dimensions, customer behavior dimensions, and cross-system feedback dimensions according to the instructions:

[0062] Basic attribute dimensions: These include basic filtering conditions such as file format, upload time, file size, and business scenario tags, used to quickly locate files that meet these basic characteristics;

[0063] Customer behavior dimension: Includes filtering conditions related to customer behavior, such as customer call frequency, consultation keywords, and interaction scenario preferences, to adapt to personalized customer needs;

[0064] Cross-system feedback dimension: This includes filtering conditions corresponding to cross-system feedback data such as the conversion rate of files viewed in the customer management system and the file push opening rate in the communication push system, which are used to filter high-quality files with high adaptability.

[0065] Step 4: Receive the user's input filtering business linkage instruction, and sequentially execute the filtering dimension self-evolution processing flow, business conflict calibration processing flow, and cross-system filtering dimension generation processing flow to realize real-time data linkage between filtering results and personal account customer communication scenarios and multiple business systems, and complete the predictive generation of filtering results and dynamic calibration of personal account business processes.

[0066] For example, the computer receives a user's input instruction to filter business linkage, such as automatically filtering suitable information when a customer inquires about a product, and sequentially executes three preset processing flows to achieve real-time linkage between the filtering results and the business scenario.

[0067] Given that existing filtering dimensions are fixed, cannot dynamically adjust priorities based on customer communication behavior, and require users to manually trigger filtering, making it impossible to anticipate customer needs in advance, resulting in low filtering efficiency and insufficient adaptability, in one possible implementation, the self-evolutionary processing flow of the filtering dimensions includes the following:

[0068] Receive instructions to collect customer communication behavior data from personal accounts, and obtain customer communication behavior data such as call frequency, consultation keywords, and interaction scenario preferences through real-time data capture, and store it in the behavior database;

[0069] The dimension priority algorithm is invoked to automatically adjust the priority order of each filtering dimension based on the customer communication behavior data in the behavior database; when high-frequency call customer data is detected, the high-value tags in the customer tags are automatically set as the default filtering conditions, and the file data containing the consultation keywords are extracted first through the keyword matching algorithm.

[0070] Receive instructions to generate predicted filtering results, call the historical communication data analysis model, predict customer needs based on historical communication data, automatically filter matching file data according to the prediction results, and display the filtered results in a floating format on the personal account interface.

[0071] For example, the computer receives a customer communication behavior data collection instruction and obtains customer communication behavior data in real time through methods such as parsing personal number call recordings and capturing interaction logs: it counts the number of calls between the customer and the personal number in the past 30 days and the duration of each call; it extracts core words mentioned by the customer in the call, such as product parameters, cooperation policies, and after-sales guarantees, through speech-to-text and keyword extraction technology; it records the customer's interaction behavior with the filtered results, such as clicking on the information link pushed in the SMS or requesting to send a certain type of file repeatedly; all of the above data is stored in the behavior database and indexed according to the format of customer identification, behavior type, and data generation time.

[0072] The computer uses a dimension priority algorithm to automatically adjust the priority of filtering dimensions based on data in the behavioral database: when it detects that a customer has made at least 5 calls in the past 30 days, i.e. a high-frequency call customer, the high cooperation level in the customer tag is automatically set as the default filtering condition; if the customer mentions after-sales guarantee keywords multiple times in the call, the weight of the consultation keyword as the filtering dimension corresponding to after-sales guarantee is automatically increased, and file data containing the keyword, such as after-sales service agreement and frequently asked questions, are extracted first.

[0073] The computer receives the instruction to generate the prediction and screening results, calls the historical communication data analysis model, and, based on the customer's historical communication records, such as the last consultation on product after-sales service or incomplete information, predicts that the customer's needs are for supplementary after-sales guarantee-related information. It automatically filters and matches the file data, such as after-sales process videos and guarantee terms documents, and displays them in a semi-transparent floating window on the personal number call interface for easy viewing or push to users.

[0074] Considering that the existing screening logic does not take into account business scenario conflicts, which affects business efficiency, in one possible implementation, the business conflict calibration process includes the following:

[0075] Receive a business scenario conflict identification instruction, call the conflict analysis algorithm, and identify the conflict type by parsing the filtering results, using feedback data and business system format compatibility data. The conflict type includes at least content adaptation conflict and format compatibility conflict.

[0076] Upon receiving the filtering logic calibration command, for content adaptation conflicts, it automatically increases the weight parameters of the corresponding filtering dimensions for video files and simultaneously generates optimized prompt data for file format conversion; for format compatibility conflicts, it calls the file content matching algorithm to filter alternative file data that is consistent with the original file content and adds format incompatibility identification information to the original file data.

[0077] The system receives instructions to generate business process optimization suggestions, and generates personal account business process optimization suggestion data based on conflict record data in the historical conflict database. After receiving instructions from the user to confirm the adoption of the optimization suggestions, the system automatically updates the format compatibility rule parameters in the filtering logic.

[0078] For example, a computer receives a business scenario conflict identification instruction, calls a conflict analysis algorithm to parse the filtering results, and uses feedback data and business system format compatible data.

[0079] The feedback data used in the screening results refers to situations where, after receiving the product technical document pushed to the customer, the document is in PDF format and the feedback from the communication push system indicates that the file cannot be opened or the manually annotated content is too technical and difficult to understand.

[0080] The business system's format compatibility data refers to a filtering result containing a product demonstration video in MP4 format, but the communication push system's hang-up SMS function only supports image and short link formats, indicating a format incompatibility.

[0081] Based on the above data, the conflict type is identified as either content adaptation conflict or format compatibility conflict.

[0082] To address content compatibility conflicts, i.e., customer feedback that documents are difficult to understand: the computer automatically increases the weight parameters of the corresponding filtering dimensions for video and text files, and generates optimization prompt data in sync. It is recommended to convert PDF documents into graphic illustrations or demonstration videos and push them to the user management backend.

[0083] To address format compatibility issues, specifically video formats that do not support SMS push notifications: The computer uses a file content matching algorithm to filter out image and text summaries from the knowledge base that match the video content. These summaries are in JPG format and are used as replacement files. The original video file is then marked with identification information indicating format incompatibility and suggesting conversion to image and text format.

[0084] The computer receives instructions to generate business process optimization suggestions. Based on multiple format compatibility conflict records stored in the historical conflict database, it generates optimization suggestion data, suggesting that the file format of the SMS push notifications be unified to either image / text or short links. After the user confirms and adopts the suggestion through the management backend, the computer automatically updates the format compatibility rule parameters in the filtering logic, and prioritizes matching files that conform to the SMS push notification format during subsequent filtering.

[0085] Considering that existing filtering dimensions do not incorporate cross-system business feedback, such as customer conversion rates and file open rates, the filtering results may meet the basic conditions but fail to achieve good business results and support business conversion. In one possible implementation, the cross-system filtering dimension generation process includes the following:

[0086] Receive feedback collection instructions from multiple business systems, establish data interaction with the customer management system and communication push system respectively, obtain customer conversion rate data after file viewing and SMS open rate data corresponding to the file, and store them in the cross-system feedback database;

[0087] Upon receiving the instruction to adjust the weight of the filtering dimension, the threshold comparison algorithm is invoked. If the customer conversion rate data corresponding to the file reaches the preset high conversion threshold, the weight value of the knowledge type dimension of the file is automatically increased; if the SMS open rate data corresponding to the file is lower than the preset effective open rate threshold, the priority parameter of the file in the idle SMS push business scenario is automatically reduced.

[0088] The system receives a cross-system filtering dimension generation instruction, calls a dimension combination algorithm, and generates new filtering dimensions based on multi-business system feedback data in the cross-system feedback database. The new filtering dimensions include at least a combination of dimensions representing high conversion and high open rate. After receiving the filtering instruction input by the user, the system completes the filtering of file data based on the new filtering dimensions.

[0089] For example, the computer receives feedback collection instructions from multiple business systems and obtains feedback data from the customer management system and the communication push system through established interfaces: From the customer management system, it obtains the customer conversion rate after viewing a file: the conversion rate of customers who reach a cooperation agreement within 3 days after viewing a product price list; From the communication push system, it obtains the file SMS open rate: the proportion of customers who click on the link to view selected product information after it is pushed via SMS; The above data is stored in a cross-system feedback database, and is associated and stored according to file identifier, conversion data, open rate data, and collection time.

[0090] The computer receives the instruction to adjust the weight of the filtering dimensions, calls the threshold comparison algorithm, and compares the feedback data corresponding to the file with the preset threshold: if the customer conversion rate of a product introduction file reaches the preset high conversion threshold, the weight value of the knowledge type dimension of the file is automatically increased, and the file will be displayed first when filtering product introduction related files in the future.

[0091] If the SMS open rate of a certain coupon file is lower than the preset effective open rate threshold, the priority parameter of the file in the idle SMS push business scenario will be automatically reduced to reduce the probability of it being selected for push.

[0092] The computer receives a cross-system filtering dimension generation instruction, calls a dimension combination algorithm, and generates a new filtering dimension based on data from the cross-system feedback database. This new dimension is a high-conversion, high-open-rate combination dimension. The filtering criteria for this dimension are that the customer conversion rate is not lower than a preset high-conversion threshold and the SMS open rate is not lower than a preset effective open rate threshold. When the user enters a filtering instruction to filter high-quality promotional materials, the computer directly filters the file data that meets the requirements based on this new dimension, without the need to manually combine multiple filtering criteria.

[0093] In one possible implementation, achieving real-time data linkage between screening results and personal account customer communication scenarios, as well as multiple business systems, includes the following:

[0094] Receive batch association instructions from users, call the scenario matching algorithm, and match and recommend suitable business scenario data based on cross-system feedback data to the filtering results;

[0095] Receive batch activation status configuration instructions, provide activation method options for the filtered results, the activation methods include at least immediate activation and activation for a specified time period; call the time period matching algorithm to automatically verify the overlap between the user's selected specified time period and the customer's active peak time period determined based on historical call data, and output the verification results to help improve the file reach rate.

[0096] In one possible implementation, the dynamic calibration of the personal number service process includes the following:

[0097] Upon receiving a business process calibration trigger instruction, the system identifies the deviation types in the personal account business process by parsing the linkage data between the filtering results and the business scenario. The deviation types include at least filtering result delivery delay deviation and business scenario adaptation deviation. The system then invokes a process optimization algorithm to generate a business process calibration plan based on the deviation types and historical linkage data. The calibration plan includes at least filtering result push timing adjustment parameters and business scenario matching rule update suggestions. After receiving a user confirmation instruction for calibration, the system automatically executes the calibration plan and synchronously stores the process parameters and business effect data before and after calibration in the process calibration database for subsequent optimization and iteration of the calibration plan.

[0098] Example 2

[0099] Based on the above embodiments, the existing file upload process lacks automatic association with customers' historical communication behavior, resulting in inaccurate customer tag allocation, affecting subsequent screening results, and file data cannot be quickly located, increasing screening time. The file upload includes the following:

[0100] The system receives file behavior association instructions from the user, calls the historical data association algorithm, automatically associates the historical customer communication behavior data of the personal account when uploading file data, and initially assigns corresponding customer tags to the file data through tag matching.

[0101] The system receives an instruction to automatically generate an index, and generates cross-system call index data based on customer tags and business scenario association attributes of file data. The cross-system call index data is used to quickly locate file data in the subsequent filtering process.

[0102] For example, when executing the file upload logic in step 2, the computer receives the file behavior association instruction input by the user, calls the historical data association algorithm, and automatically associates the historical customer communication behavior data of the personal account when uploading the Q3 2024 product quotation. It identifies that three customers have inquired about the relevant product quotation within the past month, and thus initially assigns a high-intent customer matching tag to the file. At the same time, it receives the instruction to call the index automatic generation, and generates cross-system call index data based on the high-intent customer matching tag and cooperation negotiation support business scenario attributes of the file. The cross-system call index data includes the file storage path, tag code, and scenario code. The file can be quickly located through this index during subsequent filtering without traversing the entire knowledge base.

[0103] Example 3

[0104] Based on the above embodiments, existing filtering rules lack template management, resulting in the need for reconfiguration for each filtering, which is inefficient and cannot automatically optimize filtering conditions according to changes in customer behavior. The multi-dimensional filtering logic also includes a filtering rule template configuration process, which includes the following:

[0105] Receive the user's input instruction to save the filter template, and store the combined logic, which includes filter conditions based on customer behavior and cross-system feedback, in the template database as a template.

[0106] Receive template self-optimization instructions, monitor customer behavior changes and cross-system feedback updates in real time, and call template adjustment algorithms to automatically update the filter condition parameters in the filter template.

[0107] For example, when constructing the multi-dimensional filtering logic in step 3, the computer receives the user's input instruction to save the filtering template. It stores the combined filtering logic, which includes customer behavior dimensions of high-intent consultation, cross-system feedback dimensions of high conversion and high open rate, and business scenario of cooperation negotiation support, in the template database under the template name "High-Quality Materials for Negotiation". Subsequently, when the user switches to the cooperation negotiation business scenario, the computer receives the template self-optimization instruction and monitors customer behavior change data in real time. For example, if high-intent customers are paying more attention to after-sales service terms recently, and if cross-system feedback update data is updated, such as an increase in the open rate of after-sales terms documents, the computer automatically adjusts the filtering condition parameters in the template, including knowledge type as after-sales terms in the template filtering conditions, without requiring manual modification by the user.

[0108] Example 4

[0109] Based on the above embodiments, considering that the existing system lacks hierarchical control of filtering permissions, which may lead to unauthorized users accessing sensitive information and pose security risks, and cannot effectively manage the filtering permissions of different roles, the multi-dimensional filtering logic also includes a hierarchical control process for filtering permissions, which includes the following:

[0110] Receive user input of role behavior permission binding instructions, assign corresponding operation permission data according to personal account roles, the personal account roles include at least creator, administrator and ordinary user, and establish a mapping relationship between each role and customer communication behavior data viewing permissions;

[0111] Upon receiving a permission conflict warning command, the system invokes a permission verification algorithm. If it detects that a role has initiated a request to filter highly sensitive files that exceed its permission scope, it automatically triggers a warning signal and intercepts the filtering request. At the same time, it generates permission request prompt data and pushes it to the user interface.

[0112] For example, during the multi-dimensional filtering logic construction in step 3, the computer receives the user's input of a role behavior permission binding instruction and assigns operation permissions according to the user's role:

[0113] The creator, or account manager, has access to all filtering dimensions and can view file data associated with all customer tags; the administrator, or team leader, has filtering permissions except for core customer tags and can view relevant files for non-core customers; the ordinary user, or intern assistant, only has filtering permissions for basic attribute dimensions and public business scenarios and cannot view file data for highly sensitive customers.

[0114] At the same time, the computer receives a permission conflict warning instruction. When a regular user attempts to filter a high-sensitivity document, such as a quotation sheet exclusive to core customers, the permission verification algorithm detects that the permission has exceeded the scope, automatically triggers a warning pop-up, and intercepts the filtering request. At the same time, it generates prompt data, informing the user that they need to apply for core customer data viewing permission, and pushes it to the regular user interface and the creator's review interface.

[0115] Example 5

[0116] Based on the above embodiments, considering that the existing business conflict calibration processing flow lacks efficient utilization of historical records, resulting in repeated analysis for each conflict handling, which is inefficient and cannot quickly retrieve historical calibration schemes, the process of generating personal number business process optimization suggestion data based on conflict record data in the historical conflict database also includes the following:

[0117] The conflict record data in the historical conflict database is classified and organized, and a classification index is established according to the conflict type, which includes content adaptation conflict and format compatibility conflict.

[0118] The calibration schemes corresponding to various conflicts, the calibrated business feedback data, and the conflict type classification index are associated and stored to form an association record. When the same type of business scenario conflict is detected again, the corresponding association record is quickly retrieved through the classification index, and new business process optimization suggestion data is generated based on the historical calibration schemes and business feedback data.

[0119] For example, during the execution of step 4 of the business conflict calibration process, the computer categorizes and organizes the conflict record data in the historical conflict database, establishing a classification index based on two conflict types: content adaptation conflict and format compatibility conflict. The calibration solutions corresponding to each type of conflict, such as PDF to text / image conversion and video to short link conversion, along with the calibrated business feedback data, such as customer acceptance and file effective usage rate, are stored in association with the classification index, forming a complete association record. When a format compatibility conflict is detected again, such as an uploaded AVI video failing to be pushed via SMS, the computer quickly retrieves the corresponding association record through the classification index. Based on the historical video-to-short link calibration solution and the business feedback data showing an 85% short link open rate, new business process optimization suggestion data is generated, recommending that the AVI video be converted to an MP4 short link to adapt to the SMS push interface, thus shortening the time required to generate the optimization suggestion and improving conflict handling efficiency.

[0120] The embodiments described above are merely illustrative of specific implementations of the present invention, and while the descriptions are detailed, they should not be construed as limiting the scope of the present invention. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of the present invention, and these modifications and improvements all fall within the scope of protection of the present invention.

Claims

1. A method for creating and configuring a personal account knowledge base that supports multi-dimensional filtering, characterized in that, Includes the following steps: Step 1: Receive the user's input instruction to associate a personal account, create a dedicated knowledge base bound to the personal account, and simultaneously establish communication connections between the personal account and the customer management data interface and the multi-business system data interface. The multi-business system includes at least the customer management system and the communication push system. Step 2: Receive user input of basic knowledge base configuration instructions, and execute the logic for uploading configuration files, displaying file lists, deleting files, and importing multi-format files. The file upload logic supports batch upload and single upload, and the multi-format files must cover at least documents, audio, video, and images. At the same time, an initial mapping relationship between files and customer tags and business scenarios is established through data association algorithms. Step 3: Receive the user's input instruction to build an intelligent filtering system, and build a multi-dimensional filtering logic that includes basic attribute dimensions, customer behavior dimensions, and cross-system feedback dimensions. The multi-dimensional filtering logic includes a preset filtering dimension self-evolution processing flow, a business conflict calibration processing flow, and a cross-system filtering dimension generation processing flow. Step 4: Receive the user's input of the filtering business linkage instruction, and execute the filtering dimension self-evolution processing flow, business conflict calibration processing flow and cross-system filtering dimension generation processing flow in sequence to realize the real-time data linkage between the filtering results and the personal account customer communication scenario and multiple business systems, and complete the predictive generation of the filtering results and the dynamic calibration of the personal account business process. The self-evolutionary processing flow for the filtering dimensions includes the following: It receives instructions to collect customer communication behavior data from personal accounts, acquires customer communication behavior data through real-time data capture, and stores it in the behavior database; customer communication behavior data includes customer call frequency, consultation keywords, and interaction scenario preferences; The dimension priority algorithm is invoked to automatically adjust the priority order of each filtering dimension based on the customer communication behavior data in the behavior database; when high-frequency call customer data is detected, the high-value tags in the customer tags are automatically set as the default filtering conditions, and the file data containing the consultation keywords are extracted first through the keyword matching algorithm. Receive instructions to generate prediction and filtering results, call the historical communication data analysis model, predict customer needs based on historical communication data, automatically filter matching file data according to the prediction results, and display the filtering results in a floating format on the personal account interface. The business conflict calibration process includes the following: Receive a business scenario conflict identification instruction, call the conflict analysis algorithm, and identify the conflict type by parsing the filtering results, using feedback data and business system format compatibility data. The conflict type includes at least content adaptation conflict and format compatibility conflict. Upon receiving the filtering logic calibration command, for content adaptation conflicts, it automatically increases the weight parameters of the corresponding filtering dimensions for video files and simultaneously generates optimized prompt data for file format conversion; for format compatibility conflicts, it calls the file content matching algorithm to filter alternative file data that is consistent with the original file content and adds format incompatibility identification information to the original file data. The system receives instructions to generate business process optimization suggestions, generates personal account business process optimization suggestion data based on conflict record data in the historical conflict database, and automatically updates the format compatibility rule parameters in the filtering logic after receiving instructions from the user to confirm the adoption of the optimization suggestions. The cross-system filtering dimension generation process includes the following: Receive feedback collection instructions from multiple business systems, establish data interaction with the customer management system and communication push system respectively, obtain customer conversion rate data after file viewing and SMS open rate data corresponding to the file, and store them in the cross-system feedback database; Upon receiving the instruction to adjust the weight of the filtering dimension, the threshold comparison algorithm is invoked. If the customer conversion rate data corresponding to the file reaches the preset high conversion threshold, the weight value of the knowledge type dimension of the file is automatically increased; if the SMS open rate data corresponding to the file is lower than the preset effective open rate threshold, the priority parameter of the file in the idle SMS push business scenario is automatically reduced. The system receives a cross-system filtering dimension generation instruction, calls a dimension combination algorithm, and generates new filtering dimensions based on multi-business system feedback data in the cross-system feedback database. The new filtering dimensions include at least a combination of dimensions representing high conversion and high open rate. After receiving the filtering instruction input by the user, the system completes the filtering of file data based on the new filtering dimensions.

2. The method for creating and configuring a personal account knowledge base that supports multi-dimensional filtering according to claim 1, characterized in that, The file upload includes the following: The system receives file behavior association instructions from the user, calls the historical data association algorithm, automatically associates the historical customer communication behavior data of the personal account when uploading file data, and initially assigns corresponding customer tags to the file data through tag matching. The system receives an instruction to automatically generate an index, and generates cross-system call index data based on customer tags and business scenario association attributes of file data. The cross-system call index data is used to quickly locate file data in the subsequent filtering process.

3. The method for creating and configuring a personal account knowledge base supporting multi-dimensional filtering according to claim 1, characterized in that, The multi-dimensional filtering logic also includes a filtering rule template configuration process, which includes the following: Receive the user's input instruction to save the filter template, and store the combined logic, which includes filter conditions based on customer behavior and cross-system feedback, in the template database as a template. Receive template self-optimization instructions, monitor customer behavior changes and cross-system feedback updates in real time, and call template adjustment algorithms to automatically update the filter condition parameters in the filter template.

4. The method for creating and configuring a personal account knowledge base that supports multi-dimensional filtering according to claim 1, characterized in that, The multi-dimensional filtering logic also includes a filtering permission hierarchical control process, which includes the following: Receive user input of role behavior permission binding instructions, assign corresponding operation permission data according to personal account roles, the personal account roles include at least creator, administrator and ordinary user, and establish a mapping relationship between each role and customer communication behavior data viewing permissions; Upon receiving a permission conflict warning command, the system invokes a permission verification algorithm. If it detects that a role has initiated a request to filter highly sensitive files that exceed its permission scope, it automatically triggers a warning signal and intercepts the filtering request. At the same time, it generates permission request prompt data and pushes it to the user interface.

5. The method for creating and configuring a personal account knowledge base supporting multi-dimensional filtering according to claim 1, characterized in that, To achieve real-time data linkage between the filtering results and customer communication scenarios via personal accounts, and across multiple business systems, the following are included: Receive batch association instructions from users, call the scenario matching algorithm, and match and recommend suitable business scenario data based on cross-system feedback data to the filtering results; Receive batch activation status configuration instructions and provide activation method options for the filtered results, wherein the activation methods include at least immediate activation and activation for a specified period of time; The time period matching algorithm is invoked to automatically verify the overlap between the user's selected time period and the customer's peak activity time period determined based on historical call data, and the verification results are output to help improve the file reach rate.

6. The method for creating and configuring a personal account knowledge base supporting multi-dimensional filtering according to claim 1, characterized in that, The dynamic calibration of the personal account business process includes the following: Upon receiving a business process calibration trigger instruction, the system identifies the deviation types in the personal account business process by parsing the linkage data between the screening results and the business scenario. The deviation types include at least screening result delivery delay deviation and business scenario adaptation deviation. The system then calls a process optimization algorithm to generate a business process calibration plan based on the deviation types and historical linkage data. The calibration plan includes at least screening result push timing adjustment parameters and business scenario matching rule update suggestions. After receiving the user's confirmation instruction for calibration, the calibration scheme is executed automatically, and the process parameters and business effect data before and after calibration are stored in the process calibration database simultaneously for subsequent optimization and iteration of the calibration scheme.

7. The method for creating and configuring a personal account knowledge base supporting multi-dimensional filtering according to claim 1, characterized in that, The process of generating personal account business process optimization suggestion data based on conflict record data in the historical conflict database also includes the following: The conflict record data in the historical conflict database is classified and organized, and a classification index is established according to the conflict type, which includes content adaptation conflict and format compatibility conflict. The calibration schemes corresponding to various conflicts, the business feedback data after calibration, and the conflict type classification index are associated and stored to form an associated record. When the same type of business scenario conflict is detected again, the corresponding related records are quickly retrieved through the classification index, and new business process optimization suggestion data is generated based on historical calibration schemes and business feedback data.

Images