Government affair-oriented AI intelligent reservation and guide system and method

Abstract

The invention relates to the technical field of government affair management, in particular to an AI intelligent appointment and guide system and method for government affairs, and the system comprises an appointment information collection module which receives an online appointment application or a field number taking signal, extracts a business type and an expected working period in a request, and builds a preliminary request record; and verifying the information integrity of the preliminary request record, distributing a queuing number and confirming a reservation time point, and generating a structured task demand. According to the invention, the service type and the time period preference during the reservation of the citizens are captured and are converted into the structured task demand, so that the service request is ensured to have clear guidance from the source. A personnel skill portrait is established, hard indexes such as business proficiency and language ability are recorded, personalized factors such as work preference and learning willingness are integrated, and a dynamic and multi-dimensional available personnel view is generated in combination with real-time online state and task load monitoring.

Description

Technical Field

[0001] The present invention relates to the field of government affairs management technology, and in particular to an AI intelligent appointment and handling guidance system and method for government affairs. Background Art

[0002] The field of government administration technology is a comprehensive field that focuses on using modern management theories, information technology, and system engineering methods to optimize public affairs management and public service provision.

[0003] The existing government administration technology field often shows certain limitations in practical operation, especially in the dispatch and resource matching of front-line services for the public. The traditional queuing or simple rotation allocation mode is often based on the principle of first-come-first-served or fixed duty arrangements. This mode of operation may lead to a mismatch between business needs and the capabilities of the personnel in charge. For example, the business that requires communication in a specific language or proficiency in a complex policy answer is randomly assigned to staff who do not have the corresponding capabilities or are inexperienced, which in turn leads to service interruptions, inefficiency, repeated consultations by citizens, and even complaints. Summary of the invention

[0004] The purpose of the present invention is to solve the shortcomings of the prior art and to propose an AI intelligent appointment and guidance system and method for government affairs.

[0005] In order to achieve the above-mentioned purpose, the present invention adopts the following technical scheme: The AI ​​intelligent appointment and guidance system for government affairs includes: The appointment information collection module receives online appointment applications or on-site number collection signals, extracts the business type and expected working hours in the request, and establishes a preliminary request record; verifies the information completeness of the preliminary request record, allocates queue numbers and confirms appointment time points, and generates structured task requirements; The staff skill profiling module accesses staff files, records staff skills, including business type proficiency, language proficiency and qualification certification, collects staff task type preferences and learning willingness statements, and generates staff capability tag sets; monitors the current online status and task load of staff, combines the employee capability tag set with the online status and task load, and generates a real-time available staff list; A rule task matching module, based on the structured task requirements and the real-time available personnel list, calls the staff skill requirements corresponding to each business type in the preset rule library, compares the skill items in the structured task requirements with the skill items of the personnel in the real-time available personnel list, generates a candidate matching personnel list, processes the candidate matching personnel list according to the staff preference settings and the working time matching degree, selects the undertaking personnel, and generates the task instructions to be assigned; The scheduling collaboration support module pushes the task information to the target staff terminal or queue based on the task instructions to be assigned, receives the internal assistance request initiated by the staff, searches for matching colleagues from the real-time available personnel list according to the skills required by the request, and forwards the assistance information.

[0006] Preferably, the steps of obtaining the preliminary request record are: Receive online appointment application or on-site number collection signal, parse the request source category, extract the business type field and expected working time field, classify the request source category, filter the available business type parameters, parse the time format of the expected working time, match the appointment time standard, unify the time format and integrate the business type field and the expected working time field to generate preliminary request information; Based on the preliminary request information, searching whether the service type field is in the list of available services, and verifying whether the expected working hours field meets the preset time range, and generating a complete request record; The complete request record is stored in a database, associated with user information, and a preliminary request record is generated.

[0007] Preferably, the steps of obtaining the structured task requirements are: Based on the preliminary request record, query the pending request queue of the current business type, calculate the queue length, obtain the current queue number range, assign a unique queue number, and generate queue information with an appointment time; Based on the queue information with appointment time, data is sorted according to business type fields, queue numbers and appointment time points to establish structured task requirements.

[0008] Preferably, the steps of obtaining the employee capability tag set are: Access the staff file system to extract each employee's business proficiency, language ability and qualification certification information, and simultaneously obtain the employee's task completion records under different business types, including the completion time, success rate and error rate of each task, to generate an employee skill data set; Based on the employee skill data set, the skill adaptation index of each employee is calculated using the following formula: ; in, is the skill adaptation index, Representing employees in The normalized skill value of each skill area, is the total number of skill areas, Representing employees in The average completion time of historical tasks, The total number of historical tasks performed for the employee, Representing employees in The error rate in the task, The total number of tasks performed for the employee, Representative The global average error rate of tasks of the same type, The total number of historical tasks of this business type; Based on each employee's skill adaptation index, combined with the task type preference and learning willingness statement recorded in the employee's file, the skill adaptation index is matched with the task type preference, and the employee's task category is screened according to the skill adaptation index and learning willingness to generate an employee capability label set.

[0009] Preferably, the steps of obtaining the real-time available personnel list are: Monitor the current online status of staff members, obtain the real-time login status of each staff member, including whether they are currently online, activity time in the last day, online time, current task list and task completion status, and generate an employee online status data set; Based on the employee online status data set, the task load index of each employee is calculated using the following formula: ; in, is the task load index, Represents the total duration of the task currently being performed by the employee. Represents the cumulative online time of the employee in the current period. Represents the total number of tasks completed by the employee in the past cycle. Represents the average workload of the current queued tasks, Represents the average workload of employees in handling tasks in the past. Represents the employee's current remaining task amount, Represents the maximum number of tasks that can be executed by the employee in the current period; Based on the task load index and the employee capability tag set, employees whose task load index is lower than a threshold and who are currently online are screened, and the capability tags are matched with the tasks to be processed to obtain a real-time available personnel list.

[0010] Preferably, the steps of obtaining the candidate matching personnel list are: Extract the skill standards required for each business type, and at the same time extract the skills required for each task from the structured task requirements, and build a demand skill matrix and a personnel skill matrix by comparing the real-time available personnel list; Based on the demand skill matrix and the personnel skill matrix, the compatibility between each staff member and the task requirements is calculated using the following formula: ; in, For the degree of fit, For the task requirements The required level of the skill, For staff in proficiency in a skill, is the total number of skill items, The number of tasks completed by the staff last month. The number of failed tasks of the staff member last month. is the task load index; Based on the degree of suitability, staff members are screened from the real-time available staff list to obtain a list of candidate matching staff members.

[0011] Preferably, the steps of obtaining the task instruction to be assigned are: Based on the candidate matching personnel list, the suitability of the staff member is calculated using the following formula: ; in, For the degree of fit, Indicates that employees scores in preferences, Indicates The preference score of the task requirement, is the total number of preference items, Indicates the time overlap between employees and tasks, Indicates the number of current tasks of the employee; Based on the degree of suitability, the candidate personnel are screened, the responsible personnel are determined, and the instructions for the tasks to be assigned are obtained.

[0012] Preferably, based on the task instruction to be assigned, the task information is pushed to the target staff terminal or queue, an internal assistance request initiated by the staff is received, and a matching colleague is searched from the real-time available staff list according to the skills required by the request, and the steps of forwarding the assistance information are: Based on the task instruction to be assigned, push it to the target staff terminal or task execution queue, and generate a task push record; Based on the task push record, the feedback status of the target staff terminal is monitored, the received internal assistance request of the staff is parsed, the skill requirements, urgency and task status of the current staff of the assistance request are extracted, and the real-time available personnel list is called to screen the assistance personnel who meet the skill requirements and generate the assistance matching result; Based on the assistance matching result, the matched assistance personnel information is integrated with the assistance request information, and an assistance task notification is pushed to the target assistance personnel, and the assistance task allocation record is updated at the same time.

[0013] The present invention provides an intelligent reservation and guidance method, comprising the following steps: Receive the user's online appointment application or on-site number collection signal, identify the business type and expected working hours, record and classify the appointment information, and generate a preliminary request record; Based on the preliminary request record, verify the integrity of the request information, assign a queue number, confirm the appointment time, and generate detailed appointment information; Access the staff profile database to record each staff member's business proficiency, language ability and qualification certification, summarize this information with the employee's task type preference and learning willingness, and generate a set of employee capability tags and a real-time available personnel list based on the current online status and task load; Based on detailed appointment information, employee capability tag sets and a real-time list of available personnel, the preset rule library is used to match personnel with tasks. Based on employee preferences and the suitability of working hours, matching employees are selected and task instructions to be assigned are generated. Based on the instructions of the tasks to be assigned, push the task information to the staff's terminal or queue, process the internal assistance requests initiated by the staff, match colleagues to provide assistance, and generate task coordination and execution status.

[0014] Compared with the prior art, the advantages and positive effects of the present invention are: In the present invention, by capturing the business type and time period preferences of citizens when making appointments and converting them into structured task requirements, it is ensured that the service request has clear guidance from the source. By establishing a skill profile of the personnel, not only hard indicators such as business proficiency and language ability are recorded, but also personalized factors such as work preferences and willingness to learn are incorporated, and a dynamic and multi-dimensional view of available personnel is generated by combining real-time online status and task load monitoring. According to the preset rule base, the structured task requirements are matched with the real-time list of available personnel, and the skills required for the service are automatically compared with the ability labels possessed by the personnel. After the candidate list is generated, it is further screened with the help of the staff's preference settings and the fit of the working hours, so as to achieve precise and personalized personnel assignment. This matching mechanism ensures that the person in charge has the qualifications and immediate ability to handle specific business, thereby improving the success rate and quality of a single service. Task instructions can be automatically pushed to target personnel. At the same time, when staff encounter difficulties, they can quickly locate and request assistance from other internal colleagues with corresponding capabilities based on the required skills, thereby accelerating the handling of complex problems, promoting internal knowledge flow and team collaboration efficiency, and jointly working to improve the overall response speed, professional level and citizen satisfaction of government services, and achieve optimal allocation of human resources and potential discovery. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] Figure 1 It is a system flow chart of the present invention. DETAILED DESCRIPTION

[0016] In order to make the purpose, technical solution and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0017] See also Figure 1 The present invention provides a technical solution: an AI intelligent appointment and guidance system for government affairs, including: The appointment information collection module receives online appointment applications or on-site number collection signals, extracts the business type and expected working hours in the request, and establishes a preliminary request record; verifies the information completeness of the preliminary request record, allocates queue numbers and confirms appointment time points, and generates structured task requirements; The staff skill profiling module accesses staff files, records staff skills, including business type proficiency, language proficiency, and qualification certification, collects staff task type preferences and learning willingness statements, and generates staff capability tag sets; monitors staff current online status and task load, combines staff capability tag sets with online status and task load, and generates a real-time available staff list; The rule task matching module, based on the structured task requirements and the real-time available personnel list, calls the staff skill requirements corresponding to each business type in the preset rule library, compares the skill items in the structured task requirements with the skills items of the personnel in the real-time available personnel list, generates a list of candidate matching personnel, processes the list of candidate matching personnel according to the staff preference settings and the matching degree of the working hours, selects the responsible personnel, and generates the task instructions to be assigned; The scheduling collaboration support module pushes task information to the target staff terminal or queue based on the task instructions to be assigned, receives internal assistance requests initiated by staff, searches for matching colleagues from the real-time available personnel list based on the skills required by the request, and forwards the assistance information.

[0018] The steps to obtain the initial request record are: Receive online appointment application or on-site number collection signal, parse the request source category, extract the business type field and expected working time field, classify the request source category, filter the available business type parameters, parse the time format of the expected working time, match the appointment time standard, unify the time format and integrate the business type field and the expected working time field to generate preliminary request information; Based on the preliminary request information, check whether the business type field is in the list of available businesses, and verify whether the expected working hours field meets the preset time range, and generate a complete request record; The complete request record is stored in the database, associated with the user information, and a preliminary request record is generated.

[0019] Specifically, based on the request information submitted online and offline, the received source categories are classified and summarized item by item, so as to distinguish specific categories such as "online submission by individual users", "online submission by corporate users", "on-site number collection by individual users", and "on-site number collection by corporate users". Then, the available business type parameters are screened out from these categories. For example, if there are "social security", "real estate registration", "tax processing" and other processable items, the business type codes are matched with the source categories one by one, and then the format of the expected working hours is parsed. If the obtained time period string is "09:30-10:30", it is determined whether it meets the working range set in advance, for example, the daily working hours are divided into to , by comparing whether the start and end points of the time period fall within the interval to determine its validity. If the start or end of the time period exceeds this interval, it is recorded as an abnormal entry and tried to be compared with the overtime range again. If it has not exceeded the overtime limit (assuming that the additional overtime limit is allowed), to As the scope of application), the overtime application will be screened again. If it still exceeds the limit, it will be regarded as an invalid request. The specific overtime threshold and the judgment standard can be empirically fitted and set according to the statistical results of workload in previous years. For example, according to the statistics of the past year, the average number of business sessions that can be handled after get off work is 5 per week. On this basis, a certain amount of surplus can be reserved to formulate the overtime threshold and the judgment standard can be set according to the statistical results of workload in previous years. The extension value of is obtained, and the standard of no more than 7 overtime sessions per week is obtained. Based on this, it is determined whether each time period information is within this range, and then the time periods that meet the requirements are uniformly converted into a 24-hour system, such as "09{:}30-10{:}30", and the business type field is associated with the expected working time field and packaged, integrated in the form of key-value pairs and marked as "pending confirmation items". Subsequently, by comparing this item with the established time period comparison table, the specific minute segments that can be booked can be further subdivided. This comparison table is compiled according to the standard acceptance time of each business and the minimum reservable unit (such as 15 minutes or 30 minutes). At this time, the structure after screening is uniformly stored as "organized category", and finally the preliminary request information is obtained after completing the above steps in the entire execution process.

[0020] Based on the preliminary request information obtained earlier, first read the business type set from the existing business list and perform a single key match with the business type field in the preliminary information. Assuming that the business list contains items such as "1001: Social Security", "1002: Real Estate Registration", and "1003: Tax Processing", compare the business type code in the preliminary request information accordingly. If there is a business type field with a value of "1002" and the code is in the list, it means that it belongs to a processable business type. Otherwise, mark the record as an invalid request and terminate subsequent processing. Then further verify whether the expected working hours are in line with the preset time range set earlier, for example, determine whether "09{:}30-10{:}30" falls within to Or whether the additional condition of no more than 7 overtime sessions per week is met, the remaining sessions of the available overtime period can be queried through the above-mentioned overtime determination results. If the remaining number of sessions is 0, it is determined that the period does not meet the preset standards. The comparison thresholds or comparison standards of all these retrieval and verification operations are obtained by combining empirical data with statistical history. For example, according to the historical acceptance peak, "09{:}00-11{:}30" is judged to be a busy period. At this time, no additional overtime quota will be added. For "17{:}30-19{:}00" as a non-busy period, the number of overtime hours can be appropriately relaxed. If all business type matching and time period range verification conditions are met, a new complete record is created for the request entry, and the complete entry is stored in the intermediate data set that can be processed later, thereby generating a complete request record.

[0021] After storing the complete request record generated above into the database, in order to associate the user information, it is necessary to combine the user identification data collected above. For example, assuming that the user identification data includes an ID number or a unified social credit code, etc., then compare the identification with the account information stored in the database. If it is found that the user has registered in the system, the user identification is directly associated with the complete request record. Otherwise, check whether the most basic field requirements are met based on the unit or personal information. For example, if the unit code length needs to meet Bit to The personal ID number must meet If the verification is not satisfied, it will be recorded as "user information to be verified" in the system. If it is satisfied, a new user information data will be created and bound to the complete request record. The threshold settings for the length and format of each identity field here are based on relevant policy regulations and statistical analysis of historical information in existing databases. For example, the common length distribution of unified social credit codes of enterprises and institutions registered in the past year is: bits and no missing bits, according to this conclusion, the fixed recognition length is bits, when the length of the code submitted by the user is not It is marked as abnormal, and then the verification code is compared to see if it is correct. It is considered a valid record only after confirmation. After all comparisons are completed and associations are established, the preliminary request record is obtained.

[0022] The steps to obtain structured task requirements are: Based on the preliminary request record, query the pending request queue of the current business type, calculate the queue length, obtain the current queue number range, assign a unique queue number, and generate queue information with appointment time; Based on the queue information with appointment time, data is sorted according to business type fields, queue numbers and appointment time points to establish structured task requirements.

[0023] Specifically, based on the previously registered preliminary request record, first read the business type specified in the record and search the corresponding pending request queue for the business type in the system, and then perform numerical calculations on the current length of the queue. For example, first read the number of existing data entries in the queue and compare it with the upper limit of the daily accumulation allowed. If the queue length exceeds 200, it is considered to have reached preliminary saturation. The 200 is obtained by combining the average application volume and the peak application volume in the past two years through weighted calculation. For example, the daily peak of 250 and the average of 150 are weighted respectively. and , we can conclude The mixed standard is used to determine the queue threshold during actual execution. If the threshold is exceeded, the new request will be postponed, and the requests that meet the acceptance conditions will continue to be assigned queue numbers at the back of the queue. The maximum queue number currently generated for this business type is found by querying the existing records and added by 1 as the unique queue number for this request. Then, the remaining minutes that can be reserved on the day are determined according to the date and time period information previously specified by the user. From the statistics of working hours in previous years, it is determined that a total of 15-minute time periods can be divided from 0:00 to 24:00 in a day and they are distinguished by date identifiers. If there is an available position for the current queue number in the time period, the position is bound to the queue number and packaged to generate a queue entry with a reservation time. If the time period is fully filled, the queue number and time of this request are recorded as overloaded and enter the subsequent logical link for further processing. After completing the above process, the queue information with the reservation time is obtained.

[0024] Based on the queue information with appointment time obtained above, the service type field is matched with the queue number mentioned above and the data is sorted in combination with the appointment time point. In order to make the sorting process have a clear range judgment, a corresponding table of service classification and time period needs to be set in the system. For example, when the service type is "social insurance processing", only the service type is allowed to be processed in to When the business type is "real estate registration", the scope can be expanded to to At the same time, the queue numbers are rearranged in order from small to large to check whether there are repeated or missing numbers. If repeated numbers are found, the duplicate numbers are automatically filled in according to the unique number allocation logic shown above and adjusted to the next valid number segment. For example, when the number segment is 101 to 999 and 199 has been used, the subsequent numbers are incremented to 200, 201, etc. If an empty number is found, it is automatically skipped and the sequential accumulation continues. In this way, each request can correspond to an accurate number. Then this part of information is associated with the appointment time of different dates and it is confirmed whether the queue number limit set for the same business type has been inserted during the effective period of the day. If it exceeds, it will be postponed according to the specific upper limit of the business type. The upper limit can be determined based on the actual peak period statistics in the past six months or a year. For example, the middle value of 50 between the peak amount of 80 and the minimum amount of 20 in a month is taken as the starting point of the segmentation, and then the peak amount is weighted by 0.7 to obtain an upper limit value of about 56. After completing all checks and matching, a structured task requirement is established.

[0025] The steps to obtain the employee capability tag set are: Access the staff file system to extract each employee's business proficiency, language ability and qualification certification information, and simultaneously obtain the employee's task completion records under different business types, including the completion time, success rate and error rate of each task, to generate an employee skill data set; Based on the employee skill data set, the skill adaptation index of each employee is calculated using the following formula: ; in, is the skill adaptation index, Representing employees in The normalized skill value of each skill area, is the total number of skill areas, Representing employees in The average completion time of historical tasks, The total number of historical tasks performed for the employee, Representing employees in The error rate in the task, The total number of tasks performed for the employee, Representative The global average error rate of tasks of the same type, The total number of historical tasks of this business type; Based on each employee's skill adaptation index, combined with the task type preference and learning willingness statement recorded in the employee's file, the skill adaptation index is matched with the task type preference, and the employee's task category is screened according to the skill adaptation index and learning willingness to generate an employee capability label set.

[0026] Specifically, the staff archive system is accessed to read the basic information and previous work data of each employee and confirm their completeness. For example, business proficiency can be recorded as a distribution with a numerical range of 0 to 100, language ability can be quantitatively scored for oral expression and written expression respectively and a comprehensive value can be taken, and the qualification certification information is uniformly mapped to the internal coding sequence according to the professional qualification number obtained. At this time, it is necessary to review the completeness and consistency of several collected fields, such as by comparing the employee number with the number list of its qualification certification one by one. If the number is missing or the field is empty, it is recorded as an exception and re-inspected. After confirming that this information is in line with the pre-established legal range, the subsequent steps are entered, and then the historical tasks completed by the employee under different business types are expanded one by one, and the start time and end time of each task are counted to calculate the completion time. The time unit is minutes to ensure accuracy, and the success rate and error rate need to be collected separately. For example, in the same type of task, whether the information submitted by the employee is complete and whether the processing link is in accordance with The time closure and whether the final result data meets the predetermined business rules are checked by recording the ratio of the number of successes to the total number of times between 0 and 1 and accurate to three decimal places. Similarly, the error rate is counted between 0 and 1 and compared with the historical average error rate. If the error frequency or severity exceeds a certain predetermined statistical value, such as 0.05, it is marked. The judgment of 0.05 comes from the overall error distribution analysis of all employees in the past year. During the analysis process, the average value of the task error rate of 0.03 and the variance of 0.01 can be calculated from more than 10,000 historical data. Then, the average value plus twice the variance is used to get 0.05 as the dividing line. Subsequently, various information is classified and aggregated according to fields such as employee number, business type, completion time, success rate and error rate and put into a comprehensive table. The row records of this table correspond to the execution performance of a single employee in a specific business scenario. The value of each field matches the preset value range one by one. If it exceeds the limit, it is marked. Finally, all the merged data are summarized into an employee skill data set.

[0027] formula: The benefit of the formula is that by incorporating employees’ numerical performance in multiple skill areas, the average completion time in historical tasks, and the difference with the global error rate into the calculation, the final skill adaptation index can take into account the comprehensive skill level, work efficiency, and risk deviation.

[0028] : Employees in The normalized skill value of a skill area is usually quantified by the employee's relevant score in the skill area, such as the training time the employee has participated in in the field, the level of qualification certificates obtained, and the task performance in the corresponding position in the past six months. The original scores of multiple dimensions are added and then interval normalized to obtain the value, using the following formula: ,in It is The original cumulative score of each skill. and are the lowest and highest cumulative values ​​of the skill in all employee data. Through such a set of quantitative statistics, The score is stabilized between 0 and 100, and then it is translated or scaled to the range of 0 to 1. The specific operation can be seen in the following example: When the original score of an employee in the skill of "official document writing" is 80, and the lowest score of the same group of employees in this skill dimension is 30 and the highest score is 90, then , and further divided by 100 to become 0.8333 as the final .

[0029] : The total number of skill areas, which indicates the number of skill dimensions taken into consideration in the current calculation. This value is obtained directly through list counting. For example, when counting, business proficiency is divided into four categories: "official document writing", "policy interpretation", "business operation" and "risk prediction", then qk=4.

[0030] : Employees in The average completion time of historical tasks is obtained after multiple statistics of tasks of a specific type or similar difficulty. For tasks of the same type, the number of times an employee performs is assumed to be , you can add up the minutes it takes to complete each task and divide it by Get the average time of a single task, and then perform weighted average of tasks of different types or with high similarity. The weighted proportions need to be clearly written out. For example, if the same type of tasks are divided into 3 levels according to difficulty, tasks of difficulty level 3 are counted as weight 3, and tasks of difficulty level 1 are counted as weight 1. More consideration is given to the time impact of complex tasks. The example calculation process can be written as ,in It is The time it takes to execute the task, It is the weight given to the task according to its difficulty. After completing the statistics of all similar tasks, a stable .

[0031] :The total number of historical tasks performed by an employee refers to the number of tasks completed by the employee in a certain assessment period or data statistics period. To avoid excessive data fluctuations, the records of the past 6 months are taken. For example, after summarizing the completed tasks of an employee under all business types in the past 6 months, it is found that the employee has performed a total of 180 tasks. .

[0032] : Employees in The error rate in a task is measured by the ratio of "the number of errors or their severity" to "the total number of operations in the task" or "the total number of key operations". To make the value more stable, the operation item counting method can be introduced to split the task into several quantifiable steps. Whether an error occurs in each step is recorded as 1 or 0. Then the results of all steps are added to get the total number of errors and divided by the total number of steps to get Each task gets a specific value between 0 and 1. If a business link is particularly complex, key operations with significant impact should be counted separately and given a higher weight than ordinary operations. The specific weighting method can be used as follows: ,in Indicates A binary flag indicating whether a step is wrong. It is the important weight of this step, and after the statistics are completed, it is merged into For example, a complex task contains 10 steps, of which 2 key steps are weighted 3 each, and the remaining 8 ordinary steps are weighted 1. When the steps where errors occur are 1 key step and 2 ordinary steps, we can get .

[0033] : The total number of tasks performed by the employee, corresponding to the number of tasks recorded under all business types. If the employee has performed 180 tasks in total, .

[0034] : No. The global average error rate of tasks of the same type is the result of an overall statistical analysis of the error rates of all employees on tasks of this type in the same industry or within the same business scope. First, list the tasks performed by all employees in this business type and calculate the error rate of each task to form a benchmark error indicator for this business type. If there are 2,000 task records in a business type and all error rates are added up to get about 560, then divided by 2,000 to get a global average error rate of about 0.28, then this 0.28 can be recorded as the global average error rate of tasks of the same type. .

[0035] : The total number of historical tasks of this business type, which is used as the denominator when calculating the global average error rate. It is also used to ensure that each error rate has equal importance when making comparisons. Usually, this value is updated every time a new batch of task records of the same type are processed. If a total of 15,000 completed tasks of this type are found during a statistical period, then .

[0036] Calculation process: The first step is to calculate the normalized values ​​of employee A in the four skill areas. ,at this time ; The second step is to calculate the number of times A has performed in the past 6 months. similar or highly similar tasks, count the time it takes to complete each task, and get the average time it takes for the employee to complete these tasks Minutes, the mean is taken as 133.33; The third step is to count the employee’s error rate sequence And add and divide Get the average error rate. For example, if the sum of the error rates in 180 tasks is 45, then , and then consult the global average error rate list of the same type of tasks, assuming that the summary result is , then the difference ; Step 4: Add the skills part ; Step 5: Square the average completion time , plus the difference between step 4 and step 3 to get: ; Step 6: Take the square root .

[0037] The result shows that the skill adaptation index of employee A is 133.28. If compared with employees of the same batch, further linear normalization can be performed or comparison can be performed to see whose QD is larger or smaller. If the QD value exceeds 150, it means that the comprehensive skills and efficiency levels are significantly higher than the normal level. If the value is lower than 80, it means that there are still large gaps in certain skill dimensions and efficiency indicators.

[0038] Based on each employee's skill adaptation index, it is matched with the task type preference and learning willingness statement recorded in the employee's file. In order to achieve more refined screening in the matching process, it is necessary to first read the values ​​of the employee's skill adaptation index in the previous statistical period and distinguish them. For example, it is set to divide into multiple segments between 0 and 200, and define the skill adaptation index less than 70 as the primary segment, 70 to 120 as the intermediate segment, and 120 and above as the advanced segment. Then, the system is searched for each employee's corresponding preference items and learning willingness descriptions. The preference items can be combined with the business types that employees have actively selected in the past to form a multi-dimensional field, such as "prefer to handle real estate" and "prefer to communicate in foreign languages". Each preference can also correspond to a specific matching weight value to characterize the employee's emphasis on the preference. For example, if an employee has actively selected in the past year, If 30 foreign language communication services are selected, the matching weight can be calculated as 0.75 based on the proportion of the number of times. When the cumulative number of times is less than 5 times, it can be defined as 0.2 and recorded in the weight list. For the learning intention statement part, the text can be parsed according to the learning direction filled in by the employee and mapped to the corresponding business type entry. Then, according to the aforementioned matching weight algorithm, the weights of these learning directions are multiplied by the skill adaptation index. If the product is higher than 100, it is marked as a priority recommended business, otherwise it is a general or secondary recommendation. Through this step, each employee is associated with multiple business categories and scored, and then a list of task categories with scores from high to low are arranged respectively, and these lists are recorded one by one against the business library. If different categories with the same score level appear, they are retained in the candidate items at the same time. Finally, after this fine matching and summary, the employee capability label set is obtained.

[0039] The steps to obtain the real-time available personnel list are: Monitor the current online status of staff members, obtain the real-time login status of each staff member, including whether they are currently online, activity time in the last day, online time, current task list and task completion status, and generate an employee online status data set; Based on the employee online status data set, the task load index of each employee is calculated using the following formula: ; in, is the task load index, Represents the total duration of the task currently being performed by the employee. Represents the cumulative online time of the employee in the current period. Represents the total number of tasks completed by the employee in the past cycle. Represents the average workload of the current queued tasks, Represents the average workload of employees in handling tasks in the past. Represents the current remaining amount of tasks for the employee, Represents the maximum number of tasks that can be executed by the employee in the current period; Based on the task load index and the employee capability label set, the employees whose task load index is lower than the threshold and who are currently online are screened, and the capability labels are matched with the tasks to be processed to obtain a real-time list of available personnel.

[0040] Specifically, the current online status of the staff is monitored, and the login information of each staff member is extracted from various usage records from the previous day to the current period, and the occurrence time of the login event is checked one by one. Then, the current online information is matched with the activity time of the most recent day. If it is detected that the staff member has had no less than 3 system access events after 08:00 on the same day, it is determined that the activity frequency is high. If the cumulative online time exceeds 10 hours, it is marked as a continuous online state and its activity curve is recorded. If it is found that there are only 1 or 2 access events on the same day and the time is more than 4 hours away from the current time, it is marked as an intermittent online state. If there is no activity for more than 6 hours and no offline operation is performed, this situation is counted as a potential offline interval. The 3 times, 10 hours, 4 hours and 6 hours involved here are all from the analysis results of the management department on the working hours and usage frequency in the past year: after the statistical workday access records reach 20,000, the interval from each staff member's login to the completion of related tasks is evaluated and the inactive time period is clustered. By finding the high-frequency time threshold points, a group of values ​​are summarized for subsequent judgment, and then the time threshold values ​​of each staff member are obtained. How many tasks have been completed by the personnel on the day and at what nodes these tasks are, for example, to determine whether all information has been submitted or is still being processed. The completion status is clarified by comparing the task identifier with the timestamp. If the identifier of a task appears repeatedly in the past 2 hours but has not been changed to the completion mark, it is inferred that the employee may be processing a more complex business and it takes a long time. It is necessary to record the task number and the number of minutes consumed for subsequent verification. For tasks that have not been completed on the day but are in processing or suspended, the task number and the number of remaining links must be recorded one by one. Then the above information including online and offline segments, cumulative total online time, the list of tasks being executed and the completion mark of each task are merged into the same structure, and index matching is uniformly performed according to the employee number. During the merging process, check whether the data is repeated or missing. If multiple duplicate records are found for the same employee in the same period, the duplicate rows are filtered out by comparing the login sequence number. If it is detected that an employee does not have any records on the day, it is marked as offline and the associated processing is skipped. Finally, all the current online status data of all employees that have been sorted out are summarized in a searchable manner to obtain the employee online status data set.

[0041] formula: The usefulness of the formula lies in the introduction of parameters such as "total duration of current task" and "accumulated online time" to measure work intensity. At the same time, it combines the difference between "total number of tasks completed in the past cycle", "average workload of current queued tasks" and "historical average workload" as well as the comparison between the remaining tasks and the maximum number of executable tasks to achieve multi-dimensional quantification of employees' current load level.

[0042] : Represents the total duration of the task currently being performed by the employee. To accurately obtain this value, it is necessary to record each task in segments within the day. First, analyze the difference between the start time and the current time for each task one by one, and sum up the duration of all tasks in execution to get the original duration value. Then, weight them according to difficulty or priority and add them up using a unified rule. If some tasks are in the pre-waiting state but have occupied the employee's working time, it is necessary to count the time period actually invested by the employee since the start of the queue. For example, the time spent waiting for document review or process queuing should be included in the statistics to avoid omissions. The following algorithm can be used for this: ,in The number of tasks currently in progress or queued. For the The number of minutes that have been consumed by the task, is the weight coefficient in the task, which can be determined by reference to its urgency or complexity. After the collection is completed, it is summarized as For example, there are three tasks being executed, and the corresponding time consumption is 50 minutes, 30 minutes, and 10 minutes respectively. The urgent task is assigned a value of 2 and the ordinary task is assigned a value of 1. Minutes, in this case .

[0043] : represents the cumulative online time of the employee in the current period. It needs to be combined with the employee's first login time of the day and calculated to the sum of the current online interval. It is obtained by parsing the timestamps of each online and offline event and accumulating multiple segments. If the employee starts logging in at 08:30 and is online until 17:30, and there is no idle time of more than 1 hour during the entire period, then 540 minutes can be directly used as the online time. .

[0044] : represents the total number of tasks completed by an employee in a certain statistical period in the past. It is necessary to extract the entries with the task status of "completed" corresponding to the employee number in the historical database, limit them by time range and count them. To make the statistics more referenceable, the past 30 days data period can be selected. If employee number X completed 160 tasks within 30 days, then .

[0045] : represents the average workload of the current queued tasks. To view the task queues that have been assigned or to be assigned in the system at the current moment, the estimated workload of all queued tasks is averaged. The formula can be used ,in is the number of tasks currently in the queue, It is the estimated working time or complexity value of each queued task. For example, the average working time for household registration tasks is 30 minutes, the average working time for real estate processing tasks is 45 minutes, and the average working time for social security business tasks is 25 minutes. After statistics, take the average value to get , if the current queue contains 5 tasks, with working hours of 20, 40, 40, 25, and 55 respectively, then .

[0046] : represents the average workload of employees in handling tasks in the past, and Correspondingly, it is necessary to count the time spent on all tasks in the task information that the employee has completed and take the average, or to layer them according to the task type and then merge them into a single average. If the total number of tasks handled by an employee in the past two months is 120, and the total time spent on each task is 4320 minutes, then it can be calculated .

[0047] : Represents the current remaining number of tasks for the employee. You need to check the list of tasks that the employee has received but has not yet completed in the system and count their number. Assuming that an employee has received 7 tasks on the day and is still executing them, and no completed status is found, then .

[0048] : represents the maximum number of tasks that an employee can perform in the current time period. The upper limit needs to be set in combination with the system scheduling and business regulations. First, compare the employee's position and the operating mode of the current time period. If the position is a regular job, the range can be set between 5 and 10. If the position is highly professional and accepts complex tasks, the range can be narrowed or expanded accordingly to reflect how many tasks the employee can handle in the same time period. If statistics show that an employee in the real estate processing position has a maximum of 8 concurrent tasks, the baseline value of 7 can be taken and the value can be slightly increased to 8 considering that the employee has a long experience. Finally, 8 is recorded as this person's number. .

[0049] Calculation process: The first step is to collect all the parameters and then and First, calculate the total time that an employee is in the task execution state at this time , cumulative online time ,but ; The second step is and Perform the operation, assuming , and from the statistics of queued tasks and completed tasks, we can get and , then the difference ,at this time ; In the third step, multiply the results of the first and second steps: ; Step 4: Perform calculations, if the employee's current remaining , the maximum number of executable tasks , then the denominator , ; Step 5. Add the results of step 3 and step 4: ; From this we get ; The result shows that the employee's task load index in the current time period is 41.7778. If the data of multiple employees are calculated and compared at the same time, it is possible to identify which employees have high or low loads. For example, if an employee's H exceeds 60, it means that his or her ongoing task duration, historical task completion volume and remaining task quantity are all at a high level. If H is lower than 20, it means that the employee currently has few tasks to be processed or being processed and cannot fill the pre-set workload. When the H values ​​of multiple employees are concentrated between 40 and 50, it means that the overall load is relatively uniform. Comprehensive scheduling can be carried out based on this indicator and subsequent employee capability labels.

[0050] Based on the task load index and employee capability label set obtained previously, screen out the personnel whose task load index does not exceed 30 in the current period and who are marked as online, and evaluate the degree of connection between their capability labels and pending tasks for these personnel in turn. At this time, first read the task load index of each employee and compare it with 30. The 30 is determined by the team's selection of the boundary value between peak load and medium and low load in one year's statistics. The specific method is to collect the load index of 100 employees at different times and summarize them into 15,000 records, and then find the concentrated peak segment between 40 and 70 and the low-volume distribution segment between 10 and 30. After multiple observations, 30 is finally determined as the upper limit of the low-load interval. Then, for the list of employees who meet this interval, check their capability label sets one by one and extract the label information for business scenarios such as real estate processing, social security consultation or foreign language dialogue. Compare the matching score of each label with the requirement item of the current pending task. For example, the pending task may include the review operation of the instruction document. , it is necessary to match the two ability labels of text review and policy terms interpretation and check whether the corresponding scores of employees reach above 0.6. The 0.6 score threshold is obtained by statistically analyzing the completed status in the half-year processing records: about 800 tasks with a high success rate are selected, and the evaluation values ​​of the processing personnel in this ability are correlated with the final completion quality to generate a score and quality comparison table. Then, from the comparison table, the interval with a significantly better completion rate when the score is higher than 0.6 is found, so that this 0.6 is used as the critical value for matching subsequent task requirements with employee capabilities. After the comparison is completed, employees who meet all requirements and have a score of not less than 0.6 are listed as available candidates. If an employee has a score greater than 0.8 in multiple task requirements at the same time, he / she will be prioritized. If the score is only between 0.6 and 0.8, he / she will be ranked second. If all scores are lower than 0.6, he / she will not be included in the candidate list. Finally, a list of all qualified employees is output and the corresponding mapping relationship between ability labels and requirement items is recorded to form a real-time list of available personnel.

[0051] The steps for obtaining the candidate matching list are as follows: Extract the skill standards required for each business type, and extract the skills required for each task from the structured task requirements. Compare with the real-time available personnel list to build a demand skill matrix and a personnel skill matrix. Based on the demand skill matrix and the personnel skill matrix, the suitability of each staff member to the task requirements is calculated using the following formula: ; in, For the degree of fit, For the task requirements The required level of the skill, For staff in proficiency in a skill, is the total number of skill items, The number of tasks completed by the staff last month. The number of failed tasks of the staff member last month. is the task load index; Based on the degree of suitability, staff members are screened from the real-time available staff list to obtain a list of candidate matching staff members.

[0052] Specifically, extract the skill standards required for each business type, and read the skill fields marked for each task one by one from the structured task requirements obtained earlier to confirm the level requirements for each task in skills such as "policy interpretation", "signature review", and "multilingual communication". Then, link this information with the current real-time available personnel list. First, count the skill standards of all business types and arrange them centrally, classify the skill items corresponding to each task into a requirement list, and then break down the specific level requirements for each skill item. For example, quantify "policy interpretation" into a range from 1 to 5 levels, separate and count the oral and written expressions corresponding to "multilingual communication" and mark their respective minimum scores, divide "signature review" into two sub-items: document review and electronic signature, and record the required time and accuracy. Then, perform the same skill dimension combing for each person in the real-time available personnel list, and read the employee capability labels generated previously. Collect and employee file information, map their skill proficiency to the same dimension number. If it is found that the employee's score in "multilingual communication" is 3, it corresponds to the task requirement level 3. If it reaches level 4 in "signature review", it matches the items with task requirements above level 2. Then, these skill levels are constructed by one-to-one comparison to construct the demand skill matrix and the personnel skill matrix. In the demand skill matrix, the rows correspond to the skill items of each task, and the columns correspond to the skill level requirements. The values ​​in the matrix record the demand level values. In the personnel skill matrix, the rows correspond to each employee, and the columns correspond to their proficiency scores in each skill item. The values ​​in the matrix are the actual scores of the corresponding skill items of the employee. The two matrices are synchronized in column and row coordinates. By marking the order of specific items such as "policy interpretation", "multilingual communication", and "signature review", and storing the requirement level and the corresponding employee proficiency side by side, two comparable matrices are generated after the summary is completed for subsequent calls.

[0053] formula: The benefit of the formula is that it combines the difference between the skill requirement level and the worker’s own proficiency, and further introduces the square root of the number of tasks completed last month, the logarithm of the number of task failures last month, and the current task load index, thus achieving a multi-factor comprehensive measurement of the matching degree between personnel and tasks.

[0054] :Indicates the first The required level of a skill is assigned a numerical value through the official definition of the required capabilities of the task by the statistical unit or department. It is necessary to subdivide the requirements of different industries. For example, for "policy interpretation", a range of 1 to 5 can be defined, and each level corresponds to the degree of expertise required to complete the link. When collecting these definitions, it is necessary to retrieve the latest required level of each task in the entire business system, and then summarize the complete table and maintain it. For example, some highly complex policy interpretation links will define level 3 as the basic level, level 4 as the proficient level, and level 5 as the expert level, while simple consultation responses may only define level 1 or 2. By checking the table with the current task one by one, the corresponding value of each skill item is found, and these values ​​are used as , if in a skills needs assessment, "policy interpretation" is marked as level 3, "multilingual communication" is marked as level 2, and "signature review" is marked as level 3, then they can be assigned to wait.

[0055] :Indicates that the staff is in proficiency in a skill, and Correspondingly, it is necessary to quantify the skill level of employees. By weighting and merging the actual work efficiency, accuracy, and relevant experience time in the past few months, the results are converted to the same level range, so that "expert", "skilled", and "beginner" can correspond to 5, 3, 1 and other score segments, and statistics are made for each employee, so that they can be directly used for comparison when matching tasks. For example, statistics on the training hours and task completion status of an employee in the "policy interpretation" scenario can be summarized as follows: ,in It is a proportion set according to the requirements of the management department. The study hours can range from 0 to 200 hours, the accuracy rate can range from 0 to 1, and the historical completion volume can range from 0 to 300. After the statistics are completed, the results will be projected to the interval of 1 to 5. For example, when the study hours reach 120 hours, the accuracy rate is 0.95, and the historical completion volume is 80, it can be ,get It should be noted that when the accuracy is 0.95, it is converted into a percentage and then multiplied by 100. The final value is combined to form the same dimension, and then divided by 5 to scale the result to 1 to 5. Finally, , in a specific example it may reach level 3.7, which can be recorded as level 4.

[0056] : Indicates the total number of skill items. A comprehensive analysis is conducted in the early stage to fix the skill fields that may be involved in each task, and then a list of various skill items is formed. After that, repeated skill items are merged and numbered. After the statistics are completed, the counting results are assigned to For example, the real estate management scenario may include 5 skills, and the social security management scenario may include 3 skills. When encountering a comprehensive task, the total number of skills involved can be identified from the list. If there are 8 skills required at this time, .

[0057] : Indicates the number of tasks completed by the staff member last month. It is mainly used to measure whether his / her recent work experience is rich enough. It is necessary to read the employee's task records in the past month in the database and count the entries marked as "successfully completed". For example, if someone successfully completed 80 tasks last month, then , if the other person completes 150 .

[0058] : Indicates the number of times the staff member failed to complete tasks last month. The "failure" referred to may include being returned by customers or failing internal quality inspections. It is necessary to search the status or error identifier in the task data table to count the number of unqualified tasks handled by the employee last month. Each unqualified record is counted as one failure. If an employee fails 10 tasks within 30 days, .

[0059] : represents the task load index, which can be directly referenced from the previously obtained result of the same name, by querying the previously calculated result For example, if the employee's calculation result for the current time period is 42.33, it can be directly assigned in this formula. .

[0060] Calculation process: Step 1, calculate the skill difference part , for example, there are 3 skill requirements, , corresponding to the employee's skill value ,but ; The second step is to calculate , if the employee completed 80 tasks last month ; Step 3: Calculate , when the employee failed B=10 times in the previous month, , after taking the negative sign, ; Step 4: Add the task load index ,like , then the first three parts are combined: ; Step 5: Use this result to substitute into the exponential function: , exp(47.8721) is approximately ,pass Scaling is performed with a log base of 1.5.

[0061] The results show a numerical synthesis of the skill differences between staff and task requirements, historical task performance and current load. A larger M value often means a smaller difference with task requirements and a higher workload surplus. At the same time, a sufficient number of tasks completed last month and a low number of failures can further improve the overall level of the index. If the M value is far below the median, it means that the skill differences are large, the number of failures is high, or the load is too dispersed.

[0062] Based on the previously obtained fitness, the relevant information of each person is read from the real-time available personnel list one by one and evaluated one by one to identify the match between the key skills required for the current business and the personal capabilities of the personnel. First, a comparison table is listed item by item according to the skill items required by the business to determine whether the fitness of each skill calculated previously reaches the qualified range of 1.0 to 5.0. If it is less than the lower limit of this range, the person is recorded as temporarily not meeting the requirements, and then all personnel who meet the minimum skill threshold are marked as "skills meet preliminary conditions". Subsequently, their online status and the number of queued tasks are verified for these personnel. If their online status shows that there are no less than 2 hours of available free time during the working period from 08:00 to 17:00 on the same day and the current task volume is within 3, it means that they have potential spare capacity. Then the spare capacity information is compared with the fitness. For example, an experience threshold of 30 points is set to measure the joint score of fitness and spare capacity. When a person's fitness is 25 points and the spare capacity score is 10 points, the combined calculation totals 35 points. If it is greater than 30 points, it can be included in the secondary screening list. Otherwise, they will be eliminated in this round of screening. The 30 points are obtained by the unit based on the analysis of half-year business needs and personnel distribution data. During the analysis, 300 deployment records were summarized and the average deployment score of such positions that meet the minimum skill threshold and have remaining working time was calculated. It was observed that around 30 points is the critical position for most tasks that can be completed. Then, the degree of overlap between the historical service field and the current business type of the personnel on the secondary screening list is checked one by one, and this overlap is divided into 0% to 100%. For example, if 80% of someone’s tasks in the past month are similar businesses, the overlap will be evaluated as 80% and converted into 0.8 components to be multiplied by the aforementioned comprehensive score. After a multiplication coefficient operation is performed on the person, if the result is still more than 30, he will continue to be retained, otherwise it will be marked as skills met but the field does not match. Finally, a list of all retained personnel will be generated, and each person on the list will be sorted from high to low according to the comprehensive score, and the corresponding fitness, remaining time, historical overlap and other factors will be recorded to make the screening process traceable. Then, in the recording stage, this batch of lists will be output as a list of candidate matching personnel.

[0063] The steps to obtain the task instructions to be assigned are: Based on the list of candidate matching personnel, the suitability of the staff is calculated using the following formula: ; in, For the degree of fit, Indicates that employees scores in preferences, Indicates The preference score of the task requirement, is the total number of preference items, Indicates the time overlap between employees and tasks, Indicates the number of current tasks of the employee; Based on the degree of suitability, the candidates are screened, the responsible personnel are determined, and the instructions for the tasks to be assigned are obtained.

[0064] Specifically, the formula: The benefit of the formula is that it introduces a two-way score between preference items and task requirements, and combines the time overlap between employees and tasks and the amount of tasks that employees already have on hand to form a comprehensive preference matching structure.

[0065] : Employees in If an employee indicates a high intention for outdoor investigation tasks or has been engaged in field work for many years, the score in the preference setting can be quantified by his / her field work enthusiasm and short completion cycle. The quantification process can be carried out using the following formula: ,in They are the measurement results of three types of objective indicators, such as the number of days on the field, the quality of completion of field projects, and the total number of related businesses. If the employee has 15 days on the field in the past three months and the quality of completion of all businesses has been quantified as 0.88, and the number of field businesses is 12, then the conversion and induction may result in Then calculate .

[0066] : No. The preference score of a task requirement is used to measure the demand level of the task in some special dimensions and compare it with the employee's preferences. For example, a task may require proficiency in foreign languages ​​and willingness to visit the outdoors. In this case, these two requirements can be expressed in a quantitative form. For example, the foreign language communication requirement is rated as 80 points, and the outdoor visit requirement is rated as 60 points. The comprehensive score is written as .

[0067] : The total number of preference items. For example, when scheduling a foreign language consultation task, the preference dimensions need to include "oral interaction", "translation supplement", "business trip acceptance", "online communication proficiency" and other four items. .

[0068] : Indicates the time overlap between the employee and the task, that is, the degree of match between the executable period of the task and the current idle period of the employee. It is expressed as a decimal between 0 and 1 and can be defined as follows: ,in is the total length of time that employees can be idle, is the total time period required for the task. The two are divided to get a ratio. If the length of the employee's idle time period is exactly equal to the total time period required for task execution, the overlap can be set to 1. If the idle time period is insufficient, the overlap is less than 1. It depends on the situation to see whether the task can be carried out. For example, if someone has 3 hours of free time from 09:00 to 12:00 tomorrow and the task requirement is 2.5 hours, then , the cap value 1.0 can be taken for subsequent calculations. If the statistics are within 2.5, the value is directly taken and recorded as .

[0069] : Indicates the number of current tasks of an employee. You need to retrieve the total number of tasks that have been received but not completed in the system, and add up the tasks that are pending review, in progress, or in collaboration. For example, if a person has 4 unfinished tasks, .

[0070] Calculation process: The first step is to multiply the preference items: , for example, there are 2 preferences, , then calculate , , the product of the two ; Step 2: Calculate the denominator , if the time overlap , the number of current tasks of the employee ,but , the denominator ; The third step is to calculate the formula: ; The result shows that the comprehensive score of this employee based on the current preferred scenario and task requirements is 6550.73. If all candidates are calculated and ranked in the same way, each person's priority for the task can be clarified, thus providing a reference for subsequent formal assignments. F values ​​in different ranges represent different degrees of preference matching. If the F value is between hundreds and thousands, it means that the preferences are highly overlapped. If the F value is only within tens, it means that there are large differences in the preference dimensions and the current task period arrangement.

[0071] Candidates are screened according to their suitability. First, the suitability value of each person is read from the previously summarized candidate information and compared with the pre-established reference range. For example, the range of 30 to 60 points is considered to meet the basic conditions. Those with more than 60 points are marked as preferred recommendations, and those with less than 30 points are considered temporarily unsuitable. The reference range can be derived from the statistical results of more than 200 actual project schedulings within half a year. By recording the correlation between the suitability and the final completion degree generated after each scheduling, the average value and the dispersion are summarized to obtain the two dividing points of 30 and 60 points. Then, the qualified personnel are sorted from high to low according to the scores. If the scores are the same, they are compared with their previous The online time period or completed workload will be recorded if the workload completed within a week is significantly higher and a weighted score will be assigned. This weighted calculation can distinguish between people with the same score and list the more suitable ones. After all comparisons are completed, the first or first few people on the list will be selected as the candidate for the job, and then a final check will be made on these people with the current business type. If it is found that someone in the system has accepted two difficult tasks in the last day and the remaining time period is less than the business requirement, he or she will be eliminated or put on the standby list. At the same time, the remaining list will be output and marked as the final available contractors. Finally, a corresponding task allocation information will be generated in the data structure and the contractor's number will be marked to obtain the task instruction to be assigned.

[0072] Based on the task instructions to be assigned, the task information is pushed to the target staff terminal or queue, the internal assistance request initiated by the staff is received, and matching colleagues are found from the real-time available staff list based on the skills required by the request. The steps for forwarding the assistance information are as follows: Based on the task instructions to be assigned, push them to the target staff terminal or task execution queue, and generate task push records; Based on the task push records, the feedback status of the target staff terminal is monitored, the internal assistance requests received from the staff are parsed, the skill requirements, urgency and task status of the current staff are extracted, and the real-time available personnel list is called to screen the assisting personnel who meet the skill requirements and generate the assistance matching results; Based on the assistance matching results, the matched assistance personnel information is integrated with the assistance request information, and the assistance task notification is pushed to the target assistance personnel, and the assistance task allocation record is updated at the same time.

[0073] Specifically, based on the task instructions to be assigned, the task information is pushed to the target staff terminal or task execution queue. In order to implement the push process, it is necessary to first read the previously generated allocation information in the system background, check whether the corresponding staff number in the allocation information has an online record for the day, and whether the communication channel of the staff is unobstructed. If it is detected that the employee has interacted with the system in the past 30 minutes, it is considered to be directly connected. Otherwise, it is necessary to initiate a short message or phone call to verify. Then, after confirming that the channel is available, the task is extracted from the to-do list library, including the operation steps required for the task, the preliminary estimated duration and the urgency field, and then packaged into a transmittable data body. This data body is sent to a dedicated The system automatically generates a task push record in the format of the target interface of the staff or includes it in the queue schedule. The record retains key fields such as the push time, task number and target staff ID, and monitors whether the corresponding feedback is received within the limited time according to the previously set maximum queue time threshold, such as 15 minutes. The setting value of 15 minutes here can be based on the annual workload statistics: select the number of delayed responses exceeding 30 minutes and calculate its proportion. If it exceeds 5%, it is considered abnormal, and then divide it and select a middle duration as the new threshold. When feedback is received from the staff, the feedback time and action are written into the push record to complete the task information push.

[0074] Based on the task push record, monitor the feedback status of the target staff terminal and parse the internal assistance request received from the staff. To this end, the status changes of the task push record should be continuously tracked in the system. If it is detected that the activity of the staff identification is higher than the set standard, such as interacting with the system at least once every 5 minutes, continue to wait for their operation. If the terminal reports a request of type "internal assistance", extract the key fields in the request, including skill requirements, urgency and the task status of the current staff. The skill requirements can correspond to segmentation tags such as "legal document interpretation", "language translation" or "database query", and the urgency can be based on the request. The timeliness requirements at the time of submission are numerically divided and assigned levels. A quantile accounting for 50% is taken as the boundary from the common processing cycle distribution in historical statistics. If the urgency declared in the request is higher than this quantile, it is marked as high priority. The system then compares the current number of ongoing tasks of the staff with the required skills to form a call condition. The system then filters out employees who are not online or whose load exceeds the threshold, such as more than 10 tasks that can be accepted on the day, in accordance with the real-time list of available personnel. The system then checks the remaining personnel one by one to see if they have the corresponding skill tags and meet the urgency requirements of the request. Those who meet the requirements are listed as candidates for assistance and an assistance matching result is generated.

[0075] Based on the assistance matching results, the information of the matched assistance personnel is integrated with the assistance request information, and the assistance task notification is pushed to the target assistance personnel. In the integration stage, the personnel number, name, skill field label of the screened assistance candidates and the specific requirements of the assistance request need to be merged one by one. Since the request contains an urgency field, personnel with high-priority skills will be given priority, and these qualified employees will be compared with the remaining time of their current tasks. If the remaining time is less than the preset range, such as 30 minutes, it will be marked as temporarily unable to respond. After this comparison and filtering, 1 to 3 most suitable assistants are usually obtained, and then this part of information is included in the assistance task notification, and for example, the number of the article to be translated, the brief content of the required explanation, and the estimated period of collaboration are inserted. When the system pushes, the sending time, channel and status will be recorded and written into the assistance task allocation record. At this time, a latest confirmation time threshold is set to select the median value of 20 minutes from the collaboration completion data in the past year to determine whether the assisting party accepts the order within a reasonable time limit. If it is not confirmed at the time, the next assistance object will be searched or assigned by default, and the entire assistance task distribution process will be completed.

[0076] Based on the assistance matching results, the information of the matched assisting personnel and the assistance request information are integrated and processed and pushed to the assisting personnel who finally meet the conditions, and then the assistance task allocation record is updated. At this stage, it is necessary to determine the current online status of the assisting personnel and the amount of tasks being handled again to avoid repeated issuance of overload instructions. At the same time, it is confirmed whether the terminal of the requesting party’s staff has exited or changed the requirements during this period. If it is found that the requesting party has closed or changed the assistance requirements, the push operation is terminated and the status "terminated" is marked in the record. Otherwise, the assisting personnel information and the requesting party details are combined into an operational task description, and the text content, necessary attachments or file references are arranged in sequence in the description and the execution links are uniformly marked. The integrity and traceability of the sent content are guaranteed by timestamp encryption. Then, after the successful sending, it is recorded whether the assisting party accepts it. This operation will update the "sent" in the assistance task allocation record to "pending confirmation by the assisting party". If the assisting party confirms the order, the status is marked as "assisting". If the assisting party encounters new problems in the execution later, it can still issue further assistance requests or synchronize information with the original requesting party. After the above is completed, a complete assistance task process is formed.

Claims

1. AI intelligent appointment and guidance system for government affairs, characterized by: The system comprises: The appointment information collection module receives online appointment applications or on-site number collection signals, extracts the business type and expected working hours in the request, and establishes a preliminary request record; verifies the information completeness of the preliminary request record, allocates queue numbers and confirms appointment time points, and generates structured task requirements; The staff skill profiling module accesses staff files, records staff skills, including business type proficiency, language proficiency and qualification certification, collects staff task type preferences and learning willingness statements, and generates staff capability tag sets; monitors the current online status and task load of staff, combines the employee capability tag set with the online status and task load, and generates a real-time available staff list; A rule task matching module, based on the structured task requirements and the real-time available personnel list, calls the staff skill requirements corresponding to each business type in the preset rule library, compares the skill items in the structured task requirements with the skill items of the personnel in the real-time available personnel list, generates a candidate matching personnel list, processes the candidate matching personnel list according to the staff preference settings and the working time matching degree, selects the undertaking personnel, and generates the task instructions to be assigned; The scheduling collaboration support module pushes the task information to the target staff terminal or queue based on the task instructions to be assigned, receives the internal assistance request initiated by the staff, searches for matching colleagues from the real-time available personnel list according to the skills required by the request, and forwards the assistance information.

2. According to claim 1, the AI ​​intelligent appointment and guidance system for government affairs is characterized in that: The steps for obtaining the preliminary request record are: Receive online appointment application or on-site number collection signal, parse the request source category, extract the business type field and expected working time field, classify the request source category, filter the available business type parameters, parse the time format of the expected working time, match the appointment time standard, unify the time format and integrate the business type field and the expected working time field to generate preliminary request information; Based on the preliminary request information, searching whether the service type field is in the list of available services, and verifying whether the expected working hours field meets the preset time range, and generating a complete request record; The complete request record is stored in a database, associated with user information, and a preliminary request record is generated.

3. The AI ​​intelligent appointment and guidance system for government affairs according to claim 1 is characterized in that: The steps for obtaining the structured task requirements are as follows: Based on the preliminary request record, query the pending request queue of the current business type, calculate the queue length, obtain the current queue number range, assign a unique queue number, and generate queue information with an appointment time; Based on the queue information with appointment time, data is sorted according to business type fields, queue numbers and appointment time points to establish structured task requirements.

4. The AI ​​intelligent appointment and guidance system for government affairs according to claim 1 is characterized in that: The steps for obtaining the employee capability tag set are: Access the staff file system to extract each employee's business proficiency, language ability and qualification certification information, and simultaneously obtain the employee's task completion records under different business types, including the completion time, success rate and error rate of each task, to generate an employee skill data set; Based on the employee skill data set, the skill adaptation index of each employee is calculated using the following formula: ； in, is the skill adaptation index, Representing employees in The normalized skill value of each skill area, is the total number of skill areas, Representing employees in The average completion time of historical tasks, The total number of historical tasks performed for the employee, Representing employees in The error rate in the task, The total number of tasks performed for the employee, Representative The global average error rate of tasks of the same type, The total number of historical tasks of this business type; Based on each employee's skill adaptation index, combined with the task type preference and learning willingness statement recorded in the employee's file, the skill adaptation index is matched with the task type preference, and the employee's task category is screened according to the skill adaptation index and learning willingness to generate an employee capability label set.

5. The AI ​​intelligent appointment and guidance system for government affairs according to claim 1 is characterized in that: The steps for obtaining the real-time available personnel list are: Monitor the current online status of staff members, obtain the real-time login status of each staff member, including whether they are currently online, activity time in the last day, online time, current task list and task completion status, and generate an employee online status data set; Based on the employee online status data set, the task load index of each employee is calculated using the following formula: ； in, is the task load index, Represents the total duration of the task currently being performed by the employee. Represents the cumulative online time of the employee in the current period. Represents the total number of tasks completed by the employee in the past cycle. Represents the average workload of the current queued tasks, Represents the average workload of employees in handling tasks in the past. Represents the employee's current remaining task amount, Represents the maximum number of tasks that can be executed by the employee in the current period; Based on the task load index and the employee capability tag set, employees whose task load index is lower than a threshold and who are currently online are screened, and the capability tags are matched with the tasks to be processed to obtain a real-time available personnel list.

6. The AI ​​intelligent appointment and guidance system for government affairs according to claim 1 is characterized in that: The steps for obtaining the candidate matching personnel list are as follows: Extract the skill standards required for each business type, and at the same time extract the skills required for each task from the structured task requirements, and build a demand skill matrix and a personnel skill matrix by comparing the real-time available personnel list; Based on the demand skill matrix and the personnel skill matrix, the compatibility between each staff member and the task requirements is calculated using the following formula: ； in, For the degree of fit, For the task requirements The required level of the skill, For staff in proficiency in a skill, is the total number of skill items, The number of tasks completed by the staff last month. The number of failed tasks of the staff member last month. is the task load index; Based on the degree of suitability, staff members are screened from the real-time available staff list to obtain a list of candidate matching staff members.

7. The AI ​​intelligent appointment and guidance system for government affairs according to claim 1 is characterized in that: The steps of obtaining the task instruction to be assigned are: Based on the candidate matching personnel list, the suitability of the staff member is calculated using the following formula: ； in, For the degree of fit, Indicates that employees scores in preferences, Indicates The preference score of the task requirement, is the total number of preference items, Indicates the time overlap between employees and tasks, Indicates the number of current tasks of the employee; Based on the degree of suitability, the candidate personnel are screened, the responsible personnel are determined, and the instructions for the tasks to be assigned are obtained.

8. The AI ​​intelligent appointment and guidance system for government affairs according to claim 1 is characterized in that: Based on the task instruction to be assigned, the task information is pushed to the target staff terminal or queue, the internal assistance request initiated by the staff is received, and matching colleagues are searched from the real-time available staff list according to the skills required by the request. The steps of forwarding the assistance information are as follows: Based on the task instruction to be assigned, push it to the target staff terminal or task execution queue, and generate a task push record; Based on the task push record, the feedback status of the target staff terminal is monitored, the received internal assistance request of the staff is parsed, the skill requirements, urgency and task status of the current staff of the assistance request are extracted, and the real-time available personnel list is called to screen the assistance personnel who meet the skill requirements and generate the assistance matching result; Based on the assistance matching result, the matched assistance personnel information is integrated with the assistance request information, and an assistance task notification is pushed to the target assistance personnel, and the assistance task allocation record is updated at the same time.

9. The intelligent appointment and guidance method of the AI ​​intelligent appointment and guidance system for government affairs according to any one of claims 1 to 8, characterized in that: The following steps are involved: Receive the user's online appointment application or on-site number collection signal, identify the business type and expected working hours, record and classify the appointment information, and generate a preliminary request record; Based on the preliminary request record, verify the integrity of the request information, assign a queue number, confirm the appointment time, and generate detailed appointment information; Access the staff profile database to record each staff member's business proficiency, language ability and qualification certification, summarize this information with the employee's task type preference and learning willingness, and generate a set of employee capability tags and a real-time available personnel list based on the current online status and task load; Based on detailed appointment information, employee capability tag sets and a real-time list of available personnel, the preset rule library is used to match personnel with tasks. Based on employee preferences and the suitability of working hours, matching employees are selected and task instructions to be assigned are generated. Based on the instructions of the tasks to be assigned, push the task information to the staff's terminal or queue, process the internal assistance requests initiated by the staff, match colleagues to provide assistance, and generate task coordination and execution status.

Images