AI interaction customer service and work order full-link closed-loop management method for property scene

Abstract

This invention relates to the field of smart property services and security management technology, specifically to an AI-powered interactive customer service and work order full-link closed-loop management method for property scenarios. This method acquires service request information, generates work orders, and associates them with target service locations and target equipment assets; calculates a domain diffusion risk index and an escort reachability credibility index based on a property knowledge graph; generates a minimum risk permission gate set and temporary access credential authorization parameters, fuses them to obtain an authorization policy level, and generates escort gate constraints or a manual approval queue. During execution, it monitors and controls access control events and work order status change events. The invention fuses the domain diffusion risk index and the escort reachability credibility index to obtain the authorization policy level, escort gate constraints, or a manual approval queue, and links these to control access control events and work order status change events.

Description

Technical Field

[0001] This invention relates to the field of smart property services and security management technology, and more specifically, to an AI-interactive customer service and work order closed-loop management method for property scenarios. Background Technology

[0002] In property management service scenarios, resident or user service requests are characterized by diverse entry points, unstructured expression, and a wide range of spatial and equipment assets involved. Common entry points include mobile devices, telephone, and on-site reception. Existing solutions can typically achieve basic acceptance registration, work assignment, process feedback, and archiving. However, when service items are coupled with building spaces, access control links, critical equipment assets, and security coordination, a more refined, calculable, and traceable integrated processing method is still needed.

[0003] On the one hand, service requests are mostly submitted in natural language or via images and voice. Existing intelligent interactive customer service is often good at standard Q&A and simple triage, but in complex repair requests, cross-floor location, accurate identification of equipment assets, and extraction of service time constraints, manual data entry and verification may still be necessary. While the work order management side can record task content and processing results, the association information between work orders and target service locations and target equipment assets is often scattered in ledgers, inspection records, or historical work orders, resulting in high costs for maintaining consistency. On the other hand, work dispatching is usually matched based on personnel skills, distance, or shift schedules, but the utilization of property spatial relationships, doorway connectivity, and differences in risk of different paths is limited, making it difficult to coordinate and optimize dispatch selection and access authorization strategies within the same logical link.

[0004] Furthermore, when dealing with sensitive areas such as computer rooms, power distribution, and fire protection, temporary access authorization often employs fixed role permissions, manual door opening, or manual issuance of temporary permissions. While this meets basic access needs, there is still room for optimization in areas such as minimum authority control, path interpretability risk assessment, and consistency between authorization periods and work order service times. Security escorts are typically arranged via telephone or walkie-talkie. The actual arrival capacity, arrival uncertainty due to congestion, and the matching degree between the available escort time window and work order service times lack quantitative evaluation indicators, leading to policy diversion often relying on experience-based judgment. Moreover, during work order execution, if access control events and work order status change events lack linked verification, situations may arise where access credentials can still trigger passage after a work order is suspended or canceled, or abnormal passage lacks real-time interception and evidence chain retention. Therefore, continuous consistency verification and control of authorization parameters, minimum authority access point range, and escort constraints are needed during the execution phase. Summary of the Invention

[0005] To address the shortcomings of existing technologies, the purpose of this invention is to provide an AI-powered interactive customer service and work order closed-loop management method for property management scenarios.

[0006] To achieve the above objectives, the present invention provides the following technical solution: AI-powered interactive customer service and closed-loop management of work orders for property management scenarios include: Step 1: Obtain service request information through the interactive portal for property service recipients; Step 2: Generate structured work order data based on the service request information obtained in Step 1; generate work orders based on the structured work order data, and associate the work orders with the target service location and the target equipment asset; Step 3: Construct a property knowledge graph based on property operation data; determine a set of candidate access gates based on the target service location and target equipment assets associated with work orders, and extract them into a access influence domain subgraph and a set of access paths on the property knowledge graph; mark multi-dimensional risk weights on the access influence domain subgraph to obtain the access domain diffusion risk index; and obtain the accessibility credibility index of security escort personnel. Step 4: Generate scheduling and allocation results based on the structured data of the work order and the property knowledge graph; generate the minimum risk permission gate set and determine the authorization parameters of the temporary access certificate; obtain the escort authorization safety index by fusing the domain diffusion risk index and the escort reachability credibility index; compare the escort authorization safety index with the preset escort threshold set to determine the authorization strategy level, and generate escort gate constraints based on the authorization strategy level or transfer the work order to the manual approval queue. Step 5: During the execution of the work order, monitor and control access control events and work order status change events, and perform corresponding operations according to the authorization policy level.

[0007] In one embodiment, the property knowledge graph includes at least spatial nodes, access gate nodes, equipment asset nodes, executor nodes, and security escort personnel nodes, and includes edges representing spatial inclusion relationships, spatial connectivity relationships, access gate connection relationships, equipment asset ownership relationships, and personnel service capability relationships.

[0008] In one embodiment, the multidimensional risk weights include at least spatial type risk weights, resident density risk weights, critical facility risk weights, access point security vulnerability risk weights, and time period sensitivity risk weights.

[0009] In one embodiment, the access influence domain subgraph includes at least a set of candidate access gates, a set of spatial nodes connected to the set of candidate access gates, a public entrance spatial node, a spatial node where the target equipment asset is located, and multiple access paths from the public entrance spatial node to the spatial node where the target equipment asset is located.

[0010] In one embodiment, the domain diffusion risk index is obtained as follows: based on the multidimensional risk weights and the set of travel paths, the node risk propagation matrix is ​​calculated by performing risk propagation on the node risk in the subgraph of the travel influence domain; and the node risk propagation matrix is ​​normalized and weighted to obtain the domain diffusion risk index.

[0011] In one embodiment, the escort reachability credibility index is obtained as follows: the expected arrival time interval of the security escort personnel to each candidate access point in the candidate access point set is generated, and the arrival confidence is generated; the escortable time window interval is determined; the expected arrival time interval is matched with the escortable time window interval and the service time field in the work order structured data to calculate the time window satisfaction; the time window satisfaction and the arrival confidence are weighted and fused to obtain the escort reachability credibility index.

[0012] In one embodiment, the expected arrival time interval and arrival confidence are generated as follows: based on the real-time location data, historical trajectory data and gate congestion feature data of the security escort personnel, the expected arrival time interval of the security escort personnel to each candidate gate in the candidate gate set is calculated, and the arrival confidence is generated.

[0013] In one embodiment, the authorization strategy level includes at least a normal authorization level, an escort gate level, and a manual approval level. When the authorization strategy level is the escort gate level, an escort gate constraint is generated. When the authorization strategy level is the manual approval level, the work order is transferred to the manual approval queue.

[0014] In one embodiment, when the authorization policy level is the normal authorization level or the escort gate level, a temporary access pass is issued to the executor's terminal, and the authorization parameters of the minimum risk authority gate set and the temporary access pass are issued to the access control device; when the authorization policy level is the escort gate level, the escort gate constraints are issued to the security escort personnel's terminal.

[0015] In one embodiment, during the execution of a work order, for the received access control events and work order status change events, the access control events are checked in real time to see if they meet the authorization parameters of the minimum risk permission set and the temporary access certificate, and the accompanying gate constraints are checked when the authorization policy level is the accompanying gate level.

[0016] Compared with the prior art, the present invention has the following beneficial effects: 1) After obtaining service request information through the interaction portal, this invention further generates structured work order data and creates a work order, enabling the work order to establish a clear association with the target service location and target equipment assets; during the execution of the work order, access control events and work order status change events are monitored and controlled simultaneously, ensuring that the validity of temporary access certificates, release judgments and work order flow status are consistent, thereby improving the consistency and traceability of the entire closed loop and reducing mis-assignment, mis-access and handling deviations caused by information fragmentation.

[0017] 2) This invention constructs a property knowledge graph based on property operation data, and determines a set of candidate access gates based on the target service location and target equipment assets associated with work orders. It extracts access influence domain subgraphs and access path sets from the property knowledge graph. Multidimensional risk weights are marked on the access influence domain subgraphs to obtain the access domain diffusion risk index, and authorization parameters for the minimum risk permission gate set and temporary access certificate are generated accordingly. This ensures that the authorized gate range, access direction, number of accesses, timeout failure and service time fields are precisely aligned with the corresponding time windows, reducing risk spillover caused by excessively large authorization range or excessively wide time periods, and enhancing the interpretability and auditability of the authorization basis.

[0018] 3) This invention simultaneously obtains the accompaniment and escort reachability credibility index of security escort personnel, and integrates it with the domain diffusion risk index to form an escort authorization security index. The authorization strategy level is determined by comparing it with a preset escort threshold set. When the authorization strategy level is the escort gate level, escort gate constraints are generated. When the authorization strategy level is the manual approval level, the work order is transferred to the manual approval queue. When the authorization strategy level is the ordinary authorization level or the escort gate level, a temporary pass is issued and differentiated verification and control are performed according to the authorization strategy level during the execution phase. This achieves escort strategy classification, controllable key gate constraints, and automatic diversion of high-risk work orders, taking into account both service timeliness and security boundaries. Attached Figure Description

[0019] Figure 1 A flowchart illustrating the AI-powered interactive customer service and work order closed-loop management method for property management scenarios; Figure 2 A schematic diagram illustrating the relationship between the construction of a property knowledge graph based on property operation data and the extraction of candidate access gate sets, access influence domain subgraphs, and access path sets. Figure 3 A schematic diagram illustrating the calculation and determination of the domain diffusion risk index, escort reachability credibility index, escort authorization security index, and authorization strategy level; Figure 4 This is a flowchart illustrating the process of monitoring and controlling access control events and work order status change events during work order execution. Detailed Implementation

[0020] Reference Figures 1 to 4 AI-powered interactive customer service and closed-loop management of work orders for property management scenarios include: Step one involves acquiring service request information through the interactive entry point for property service recipients. This provides a unified business entry point and information collection starting point for AI-powered interactive customer service in property scenarios, enabling the interactive entry point for property service recipients to handle service requests from different touchpoints and aggregate them into the service request information required for subsequent processing. By acquiring service request information first, original evidence can be provided for generating structured work order data, generating work orders, and associating target service locations and target equipment assets. This avoids situations where information is missing, inconsistent, or unable to locate service recipients in subsequent stages, improving the executability and traceability of work order closed-loop management from the source.

[0021] In one specific implementation, an interactive service entry point is provided at the interaction portal for property service recipients, enabling them to initiate service requests via text or voice. Secondly, upon receiving a service request, the interaction portal prompts the property service recipient to supplement the relevant information to form a complete service request. This information may include at least target service location information, target equipment / asset information, service request time information, and the service requester's identification information. If the received information is ambiguous or incomplete, the interaction portal initiates follow-up questions based on the received content until the service request information meets the minimum information requirements for generating structured work order data. Finally, the service request information collected and verified through the interaction portal is output to subsequent steps. For example, when a property service recipient submits a voice service request for elevator noise through the interactive portal, the portal further guides them to select their building and floor as the target service location information, select the corresponding elevator number as the target equipment asset information, and confirm the expected arrival time as the service request time information, thus obtaining service request information that can be directly used for subsequent processing.

[0022] Step two involves generating structured work order data based on the service request information obtained in step one; generating work orders based on this structured data and associating them with the target service location and target equipment assets; transforming the service request information into computable, schedulable, and controllable structured work order data, and further solidifying it into a closed-loop carrier—the work order. Through this structured work order data, the system can carry the key elements of service requests in a unified field format, thereby supporting subsequent extraction, reasoning, and calculation on the property knowledge graph. By associating work orders with the target service location and target equipment assets, the system can clearly bind the destination and the task to specific spaces and assets, forming the basis for subsequent actions such as determining the candidate access point set, extracting and forming the access influence domain subgraph and access path set, generating the minimum risk permission gate set, and determining the authorization parameters of temporary access credentials. This ensures that subsequent path, risk, and permission calculations have clear objects and scopes.

[0023] In one specific implementation, the service request information undergoes element extraction and normalization processing. The service items, fault phenomena, expected response time, and constraints expressed in the service request text or voice are transcribed and organized into structured fields. Service request time-related information is standardized and written into the service time field of the work order structured data. Location parsing and consistency verification are performed on target service location-related information, mapping buildings, floors, room numbers, and public area names to unique target service location identifiers and verifying their consistency with the spatial inclusion relationship in property operation data. Asset parsing and ownership verification are performed on target equipment asset-related information, mapping equipment numbers, equipment names, and equipment categories to unique target equipment asset identifiers and verifying the matching of equipment asset ownership relationships between the target equipment asset and the target service location. Finally, a work order is generated based on the work order structured data, and the work order is associated with the target service location and target equipment asset to ensure that subsequent access point derivation, risk index calculation, and authorization strategy level determination are all based on the same associated object. For example, when a service request is for an elevator making abnormal noise on a certain floor of a certain building and requires processing before 6 PM on the same day, the processing before 6 PM is standardized into a service time field. The building and floor are parsed as the target service location, and the corresponding elevator number is parsed as the target equipment asset. The target service location and the target equipment asset are simultaneously associated in the generated work order, forming a work order data foundation that can be directly used for subsequent steps.

[0024] Step 3: Construct a property knowledge graph based on property operation data; determine a set of candidate access gates based on the target service location and target equipment assets associated with work orders, and extract them into a access influence domain subgraph and a set of access paths on the property knowledge graph; label multi-dimensional risk weights on the access influence domain subgraph to obtain the access domain diffusion risk index; obtain the accessibility and credibility index of security escort personnel; organize the property operation data into a property knowledge graph that can be used for spatial and access reasoning, and establish a calculation chain of access range—access path—risk weight—risk index on the property knowledge graph around the target service location and target equipment assets. At the same time, introduce the accessibility and credibility measurement of security escort personnel to provide a quantitative basis for subsequent authorization strategy level determination. By constructing a property knowledge graph based on property operation data, spatial nodes, access gate nodes, equipment asset nodes, executor nodes, and security escort personnel nodes, as well as their spatial inclusion relationships, spatial connectivity relationships, access gate connection relationships, equipment asset ownership relationships, and personnel service capability relationships, can be systematically expressed, providing a unified graph structure foundation for subsequent extraction and calculation. By determining the candidate access gate set based on the target service location and target equipment assets associated with work orders, and extracting them into a access influence domain subgraph and a set of access paths on the property knowledge graph, potential access entrances, access paths, and associated spatial ranges can be clearly identified. This ensures that risk assessment is no longer abstract but is grounded in a set of accessible gates and traversable paths. By labeling multidimensional risk weights on the sub-graph of the traffic influence domain, a traffic diffusion risk index is obtained. This allows the integration of factors such as spatial type risk weights, resident density risk weights, critical facility risk weights, access point security weakness risk weights, and time-sensitive risk weights into a unified risk propagation and aggregation framework. This enables the system to measure the diffusion intensity of risks within the traffic influence domain along paths and nodes, providing risk-side input for subsequent minimum risk permission thresholds, temporary access pass authorization parameters, and authorization strategy levels. Furthermore, by obtaining the escort reachability credibility index for security escort personnel, the ability of security escort personnel to reach the candidate access point set and effectively escort within the required time window can be quantified into a comparable index. This ensures that the subsequent fusion and threshold comparison of escort authorization safety indices consider not only risk but also the feasibility and credibility of escort, reducing the unfeasible nature of decisions made solely based on risk.

[0025] In one specific implementation, a property knowledge graph is constructed based on property operation data: property operation data is aggregated and fields are normalized to ensure that the same space, the same access point, the same equipment asset, the same executor, and the same security escort personnel correspond to unique identifiers in data from different sources; based on the normalization results, space nodes, access point nodes, equipment asset nodes, executor nodes, and security escort personnel nodes are generated, wherein space nodes record space type and hierarchical attributes, access point nodes record door direction and connection attributes, equipment asset nodes record asset category and location attributes, and executor nodes and security escort personnel nodes record service capability attributes; Edges are generated according to preset relationships: spatial inclusion relationship edges are generated based on spatial hierarchy; spatial connectivity relationship edges are generated based on adjacent reachability; access gate connection relationship edges are generated based on the spaces on both sides of the gate; equipment asset ownership relationship edges are generated based on the space where the asset is located; and personnel service capability relationship edges are generated based on the personnel reachable space and the types of assets that can be processed. The generated results are checked for consistency and recorded in a versioned manner to ensure that each equipment asset node corresponds to at least one equipment asset ownership relationship edge, each access gate node corresponds to at least one access gate connection relationship edge, and each personnel node corresponds to at least one personnel service capability relationship edge. Incremental updates are performed when property operation data changes. For example, the third-floor corridor and the third-floor computer room of Building 1 in a certain community are respectively regarded as spatial nodes, the access control of the third-floor corridor is regarded as the access point node, the air conditioning unit in the computer room is regarded as the equipment asset node, maintenance personnel are regarded as execution personnel nodes, and security guards are regarded as security escort personnel nodes. Then, the spatial inclusion relationship edges of Building 1 including the third floor, the third floor including the corridor and the computer room, the spatial connectivity relationship edges of the corridor and the computer room, the access point connection relationship edges of the access control connecting the corridor and the computer room, the equipment asset ownership relationship edge of the air conditioning unit belonging to the computer room, and the personnel service capability relationship edge of maintenance personnel and security guards who can serve the third-floor computer room are established respectively, thereby obtaining a property knowledge graph that satisfies subsequent reference.

[0026] In one specific implementation, the target service location and target equipment asset are read from the work order, and the spatial node where the target equipment asset is located is located in the property knowledge graph. At the same time, the spatial node corresponding to the target service location is located as a verification basis to avoid extraction deviation caused by the inconsistency between the spatial node where the target equipment asset is located and the target service location. Taking the spatial node where the target equipment asset is located as the end point and the public entrance spatial node as the starting point, a multi-path search is performed in the property knowledge graph along the spatial connectivity relationship and the access point connection relationship to obtain multiple access paths from the public entrance spatial node to the spatial node where the target equipment asset is located. During the search process, the access point nodes passed through by each access path are summarized to form a candidate access point set. The candidate access gate set, its connected spatial node set, public entrance spatial node, target equipment asset location spatial node, and multiple access paths are clipped and aggregated to form an access influence domain subgraph. This subgraph covers the spatial range that the candidate access gate set can influence while maintaining path accessibility. The access path set is deduplicated and hierarchically recorded. The gate sequence, spatial sequence, and length attributes of different paths are written into the access path set for subsequent risk weight labeling and risk index calculation. For example, when the target equipment asset associated with a work order is an air conditioning unit in a building's computer room and the public entrance space node is the community gate, the property knowledge graph searches for the access paths from the community gate to the building lobby to the elevator hall to the third-floor corridor to the computer room, as well as the access paths from the community gate to the building lobby to the stairwell to the third-floor corridor to the computer room. The elevator hall access control, stairwell access control, and computer room access control are then included in the candidate access point set. This is then used to extract and form a access influence domain subgraph that includes the candidate access point set, the set of spatial nodes connected to the candidate access point set, the public entrance space node, the space node where the target equipment asset is located, and multiple access paths.

[0027] In one specific implementation, spatial nodes, access gate nodes, target equipment asset location spatial nodes, and public entrance spatial nodes in the access impact domain sub-map are used as annotation objects. Based on property operation data, each spatial node is assigned a spatial type risk weight, and based on the number of residents and statistics of permanent and transient populations corresponding to the space, each spatial node is assigned a resident density risk weight. At the same time, spatial nodes involving key areas such as power distribution, fire protection, and weak current rooms are assigned critical facility risk weights. For each access gate node in the candidate access gate set, a security weakness risk weight is assigned based on historical abnormal door opening records, access control failure records, and inspection defect records. Meanwhile, based on the service request time-related information in the work order structured data and the time-sensitive rules in the property operation data, nodes associated with the access path set in the access impact domain sub-map are uniformly assigned time-sensitive risk weights, so that the same node has distinguishable risk labels under different service time windows. The time-sensitive rule is used to map the service time field to a time-sensitive risk weight. The mapping can adopt a segmented rule: a first time-sensitive risk weight is assigned during a preset daytime period, and a second time-sensitive risk weight higher than the first time-sensitive risk weight is assigned during a preset nighttime period; and during holidays, major event days, or temporary control periods announced by the property management, the time-sensitive risk weight is increased to a third time-sensitive risk weight, so that the same node has a distinguishable risk labeling basis under different service time windows. Based on multidimensional risk weights and a set of access paths, a node risk propagation matrix is ​​obtained by calculating the risk propagation of node risks in the subgraph of the access impact domain. The elements of the node risk propagation matrix are determined by the node's own multidimensional risk weights and its relative position in the set of access paths. The diffusion intensity of risk along the direction from the public entrance space node to the target equipment asset space node is reflected by path length attenuation and multi-path superposition. The node risk propagation matrix is ​​normalized and weighted, and different risk dimensions are aligned under a unified scale. The propagation results corresponding to the space type risk weight, resident density risk weight, key facility risk weight, access point security weakness risk weight, and time period sensitive risk weight are weighted and aggregated according to the preset dimension contribution coefficients to obtain the domain diffusion risk index used to characterize the overall diffusion risk level of the access impact domain. The node risk propagation matrix is ​​denoted as The row corresponds to the set of nodes in the subgraph of the influence domain. The column corresponds to five risk dimensions: spatial type risk weight, resident density risk weight, key facility risk weight, access point security weakness risk weight, and time period sensitivity risk weight. For any node In any risk dimension propagation value Calculate as follows: Read node In terms of risk dimension Multidimensional risk weights ,in It is a non-negative value and can be preprocessed and mapped to Determine the nodes containing the path set. Path set For each travel path Starting from the public entrance space node, the path is... Linearize and obtain the path length At the same time, the node is obtained. Along the path to the spatial node where the target equipment asset is located Remaining path length Set the path length attenuation factor. With position attenuation factor ,in and The preset attenuation coefficient is used; it is obtained by multi-path superposition. ; The node risk propagation matrix is ​​normalized column by column according to the risk dimension to obtain the normalized matrix. ,in Set the minimum positive number; set the dimensional contribution coefficient. ,satisfy and The dimensional contribution coefficients can be determined through expert configuration or through retrospective statistics based on historical abnormal door opening records, access control failure records, and inspection defect records; a weighted aggregation process is performed on the normalized matrix to obtain the domain diffusion risk index. ,in This is to enable the domain diffusion risk index to characterize the overall diffusion risk level of the domain under a unified dimension; For example, when a work order corresponds to the inspection of target equipment assets in the third-floor computer room of a building at 10 PM, the risk weight of the space type and the risk weight of the resident density in the third-floor corridor are relatively high. The computer room has a high risk weight due to its critical facilities and the access control system has a security weakness at the access points. Furthermore, the risk weight is increased due to the sensitivity of the nighttime time period. Therefore, when the set of access paths includes elevator hall paths and stairwell paths, the node risk propagation matrix will form superimposed propagation values ​​on the corridor access control, elevator hall access control or stairwell access control, computer room access control, and adjacent spatial nodes on both paths. After normalization and weighted aggregation, a higher domain diffusion risk index is obtained, which is used for subsequent authorization strategy level determination.

[0028] In one specific implementation, real-time location data of security escort personnel is read and combined with historical trajectory data to form a typical travel time distribution between different buildings and floors; for each candidate access gate in the candidate access gate set, access congestion feature data of the gate is superimposed to dynamically correct the travel time, and the expected arrival time interval is output by using quantile extrapolation, and the arrival confidence is generated by combining the update frequency of real-time location data, the stability of historical trajectory and the volatility of congestion features; the escort time window interval is determined according to the duty scheduling constraints of security escort personnel, patrol task interval and established security response rules, so that the escort time window interval can express the continuous time range that can be dispatched for escort; Based on the service time field in the structured data of the work order, the expected arrival time interval and the escort time window interval are intersected and matched to calculate the time window satisfaction. The time window satisfaction is used to characterize the degree to which the security escort personnel arrive at the corresponding candidate access point and maintain escort within the service time requirement of the work order. The time window satisfaction and arrival confidence are weighted and fused to obtain the escort reliability index of the security escort personnel. The weight can be adaptively adjusted according to the urgency of the work order service, the sensitivity of nighttime, and the degree of security weaknesses of the candidate access point. The estimated arrival time interval is generated based on statistical extrapolation of real-time location data, historical trajectory data, and gate congestion characteristics of security escorts: for any candidate gate in the candidate gate set, a set of historical travel time samples of security escorts from their current spatial node to that candidate gate is obtained based on historical trajectory data. And based on the gate traffic congestion characteristic data, the sample set was analyzed. Perform dynamic correction to obtain the corrected sample set The corrected sample set is obtained by using quantile extrapolation. lower quantile With upper quantile The estimated arrival time interval for the candidate gate is calculated by adding this time interval to the current time corresponding to the real-time location data of the security escort personnel. In one specific implementation, the arrival confidence level is used to characterize the reliability of the expected arrival time interval, assuming the update interval for the real-time location data of the security escort personnel is [missing information]. The standard deviation of the travel time for historical trajectory data is The fluctuation range of the gate traffic congestion characteristic data is Then the confidence level is reached. ,in These are preset coefficients and can be determined through expert configuration or backtesting statistics; The time window satisfaction level is used to characterize the degree to which security escort personnel arrive at the corresponding candidate access gate and maintain escort within the service time requirement of the work order. Let the service time field in the work order structured data correspond to the service time window interval as follows: The escort time window is Then the window satisfies in The length of the time interval is represented, and a preset minimum positive number can be used as a substitute when the denominator is zero; a weighted fusion of the time window satisfaction and arrival confidence is performed to obtain the escort reachability credibility index. ,in The system has preset weights and can adaptively adjust based on the urgency of the work order service, the sensitivity at night, and the security weaknesses of candidate access points. The adaptive adjustment includes increasing the weights when the urgency of the work order service is high or the sensitivity at night is high. To ensure that the satisfaction level of the time window accounts for a higher proportion in the fusion process; For example, if the service time field of the work order structured data is from 22:00 to 22:30 on the same day, and the candidate access point set includes the building lobby access control and the computer room access control, if the estimated arrival time of the security escort personnel to the building lobby access control is from 22:05 to 22:10 with a high arrival confidence, and the estimated arrival time to the computer room access control is from 22:18 to 22:35 with a medium arrival confidence, and their escort time window is from 22:00 to 23:00, then the time window satisfaction of the two access control points can be calculated separately and weighted and fused with the arrival confidence to obtain the escort reachability credibility index for subsequent escort authorization security index fusion calculation.

[0029] Step four: Generate scheduling and allocation results based on the structured data of the work order and the property knowledge graph; generate a minimum risk access gate set and determine the authorization parameters of the temporary access pass; obtain the escort authorization security index by fusing the domain diffusion risk index and the escort reachability credibility index; compare the escort authorization security index with the preset escort threshold set to determine the authorization strategy level, and generate escort gate constraints based on the authorization strategy level or transfer the work order to the manual approval queue; transform how the work order should be executed, who should execute it, where is allowed to pass, what granularity of the access authorization, whether escort is required, and whether manual approval is required into executable scheduling and allocation results and authorization control results, thereby realizing a closed loop of linkage from work order to access control authorization; By generating scheduling and allocation results based on structured work order data and property knowledge graphs, the structured elements in the work order data and the relationships between personnel service capabilities and spatial connectivity in the property knowledge graph can be used to form a feasible personnel and task allocation, ensuring that the issuance of temporary access passes and the recipients of access control devices are clearly defined. By generating a minimum risk permission gate set and determining the authorization parameters of temporary access passes, the minimum necessary access range can be defined as the gate set range, and the authorization boundaries such as available time and available gate points / paths can be defined as the authorization parameters of temporary access passes, reflecting the principles of minimum permission and risk constraints, and ensuring that access authorization is strictly matched with work order tasks. By fusing the domain diffusion risk index and the escort reachability credibility index to obtain the escort authorization security index, the domain diffusion risk index on the risk side and the escort reachability credibility index on the execution side can be unified into a comparable escort authorization security index for subsequent strategy level determination, avoiding single-index decision bias. The escort authorization security index is compared with a preset escort threshold set to determine the authorization strategy level, and escort gate constraints are generated based on the authorization strategy level, or the work order is transferred to a manual approval queue. This allows different combinations of risks and reachability credibility to be mapped to different authorization strategy levels, such as ordinary authorization level, escort gate level, and manual approval level. When the system determines that automatic authorization is possible, executable escort gate constraints are output; when the system determines that the risk or uncertainty is high, the work order is transferred to a manual approval queue to ensure security and compliance.

[0030] In one specific implementation, the service item, service time field, target service location, and target equipment asset are read from the structured data of the work order. A candidate set of executors that can cover the target equipment asset category and reach the target service location is then filtered along the personnel service capability relationship in the property knowledge graph. Simultaneously, a candidate set of security escort personnel that can cover the spatial range of the candidate access point set is also filtered along the personnel service capability relationship. The availability of the candidate set is checked for the time period based on the service time field. Indicators such as the historical handling timeliness of the executors, the success rate of handling similar faults of the target equipment assets, and the response stability of the security escort personnel in similar spaces are mapped to comparable scheduling scores. The optimal executor and the optimal security escort personnel that meet the service time field are then paired, forming a scheduling allocation result, which is written back to the work order association information. The dispatch score is obtained by weighting the historical handling timeliness of the executors, the success rate of handling similar faults of the target equipment assets, and the response stability of the security escort personnel. The historical handling timeliness can be obtained by statistically analyzing the arrival time and completion time of historical work orders. The handling success rate can be obtained by statistically analyzing whether the work order recurs or triggers repairs. The response stability can be obtained by statistically analyzing the width of the expected arrival time interval and the arrival confidence level of the security escort personnel. When any statistical item is missing, a preset default value or the average value of the same type of assets in the same building is used to fill the gap. When the filling ratio exceeds the threshold, the authorization strategy level is upgraded to the manual approval level. The scheduling and allocation results should include at least the executor's identifier. Security escort personnel identification Candidate travel path identifier Least risk access gate set The set of authorization parameters for temporary passes With authorization policy level Among them, the identification of the personnel in charge. The system filters candidates for service personnel who can cover the target equipment asset category and reach the target service location based on their service capability relationship, and then verifies the availability of personnel for different time periods based on the service time field. Security escort personnel are also identified. The candidate set of security escort personnel is selected based on the personnel service capability relationship, which can cover the spatial range of the candidate access point set, and the availability of escort personnel is verified by time period verification in combination with the available escort time window interval; candidate access route identification. The path risk value is determined under the premise of satisfying the arrival time constraint corresponding to the service time field. The minimum travel path is determined, and candidate travel paths are selected. The gate sequence is used as the connectivity benchmark for generating the gate set; the authorization parameter set It should include at least the gate access range parameters, access direction parameters, access count parameters, timeout failure parameters, and verification parameters bound to the identity of the executor, and should be consistent with the time window corresponding to the service time field; Based on the scheduling and allocation results and the set of access paths, the candidate path with the lowest overall risk and meeting the arrival time constraint is selected from multiple access paths from the public entrance space node to the space node where the target equipment asset is located. Based on the set of candidate access gates passed through by the candidate path, a minimum risk permission gate set is generated by combining risk pruning and necessary gate retention. Risk pruning is used to eliminate access gates that are only used for detours and have high risk weights, while necessary gate retention is used to ensure that the connectivity from the public entrance space node to the space node where the target equipment asset is located is not disrupted. Generating the minimum risk access gate set includes two stages: path evaluation and gate set pruning. In the path evaluation stage, a path risk value is calculated for each path based on the set of accessible paths. The path risk value can be obtained by aggregating the risk propagation matrix of the nodes traversed by the path. ,in For travel path The set of nodes traversed These are the elements of the normalized node risk propagation matrix. The contribution coefficient for the dimension; under the premise of satisfying the arrival time constraint corresponding to the service time field in the work order structured data, the travel path with the smallest path risk value is selected as the candidate path, and the gate point sequence of the candidate path is used as the connectivity benchmark for generating the gate set; During the gate set pruning phase, an initial gate set is formed based on the access gate nodes traversed by the candidate path. Risk pruning and necessary gate retention are then performed on this initial gate set: Risk pruning involves removing access gate nodes in the initial gate set whose risk weights exceed a preset gate removal threshold but do not affect the connectivity from the public entrance space node to the target equipment asset's space node; Necessary gate retention involves reversing the removal if it results in the public entrance space node becoming unreachable from the target equipment asset's space node on the access influence domain subgraph. This pruning process is repeated until no access gate nodes meet the pruning criteria, resulting in the gate set with the minimum risk authority. The preset gate removal threshold can be set based on the historical distribution of access gate security weakness risk weights and critical facility risk weights, and is updated synchronously when property operation data is updated. The effective and expiration times of temporary access passes are determined based on the service time field, and authorization parameters are determined in conjunction with the minimum risk access gate set. The authorization parameters include at least the gate access range parameter, access direction parameter, access number parameter, timeout expiration parameter, and verification parameter bound to the executor's identity identifier, so as to ensure that the temporary access pass is only effective within the time window corresponding to the work order service time field and only for the gate corresponding to the minimum risk access gate set. The escort authorization safety index is obtained by weighting and synthesizing the domain diffusion risk index and the escort reachability credibility index according to a preset fusion rule. The contribution weight of the domain diffusion risk index is increased when the proportion of space in key facilities increases at night, and the mitigation weight of the escort reachability credibility index is increased when the confidence of the security escort personnel's arrival increases and the time window is sufficiently satisfied. The escort authorization safety index is compared with a preset escort threshold set to determine the authorization strategy level. When the escort authorization safety index is lower than the ordinary authorization level threshold, it is determined to be the ordinary authorization level. When the escort authorization safety index is within the escort gate level threshold range, it is determined to be the escort gate level and escort gate constraints are generated. The escort gate constraints include at least the parameters that the security escort personnel must arrive at the key passage point first, the double arrival confirmation parameters, the passage sequence consistency parameters, and the time coupling parameters. When the escort authorization safety index is higher than the manual approval level threshold, it is determined to be the manual approval level and the work order is transferred to the manual approval queue. The critical access gate arrival parameter is used to determine the set of critical access gates that require prior confirmation by security escort personnel. The set of critical access gates is selected from the minimum risk authority gate set: the access gate nodes in the minimum risk authority gate set are sorted according to the weighted sum of the access gate security weakness risk weight and the critical facility risk weight, and at least one access gate node with the highest ranking is selected as the critical access gate; when the minimum risk authority gate set includes the last access gate node for entering the spatial node where the target equipment asset is located, the last access gate node is also included in the critical access gate set, and the gate identifier list and gate order information corresponding to the critical access gate set are recorded in the escort gate constraint conditions; The dual-person arrival confirmation parameter limits the triggering conditions for security escort personnel arrival confirmation events and execution personnel arrival events: The security escort personnel arrival confirmation event is generated by the security escort personnel terminal at a key access gate, and includes at least the key access gate identifier, the security escort personnel's identification, and the confirmation time; the execution personnel arrival event is generated by the execution personnel terminal or access control device, and includes at least the key access gate identifier, the execution personnel's identification, and the event occurrence time; dual-person arrival confirmation is deemed satisfied when both types of events occur within the time window corresponding to the service time field. The access sequence consistency parameter limits the gate order of access control events in the event timeline to be consistent with the gate order of candidate paths in the access path set. The time coupling parameter limits the maximum allowable interval between the security escort personnel confirmation time and the execution personnel's ID swiping time. Settings, in which The risk weights for security vulnerabilities at key access points, or their normalized aggregate values. and The time coupling parameter is set as a preset coefficient. When the difference between the time the executor swipes his / her ID and the time the security escort confirms his / her ID is not greater than the time coupling parameter, the time coupling parameter is deemed to be satisfied. Otherwise, the escort gate constraint condition is deemed not satisfied and a follow-up or secondary confirmation request is triggered. The escort authorization security index is used to characterize the level of risk and uncertainty in escort authorization. A higher index value indicates higher risk or uncertainty, corresponding to a higher level of authorization strategy. To facilitate standardized measurement, the domain diffusion risk index can be normalized to... get The credibility index of escorting and protecting can be normalized to... get And calculate the escort authorization security index according to the preset fusion rules. ,in Preset integration weight; increase it at night or when the proportion of space occupied by key facilities increases. To enhance the contribution of the domain diffusion risk index, it is reduced when the confidence level of security escort personnel arrival increases and the time window satisfaction is sufficient. To enhance the credible effect of escort and protection; The preset escort threshold set includes at least the ordinary authorization level threshold and the manual approval level threshold, and satisfies the condition that the ordinary authorization level threshold is less than the manual approval level threshold. When the escort authorization security index is lower than the ordinary authorization level threshold, it is determined to be at the ordinary authorization level; when the escort authorization security index is within the escort gate level threshold range, it is determined to be at the escort gate level; and when the escort authorization security index is higher than or equal to the manual approval level threshold, it is determined to be at the manual approval level. The ordinary authorization level threshold and the manual approval level threshold can be determined by retrospectively analyzing historical work orders, historical abnormal door opening records, and access control events: samples from historical work orders that underwent manual approval or experienced abnormal access are considered high-risk samples, and samples from historical work orders without abnormal closed loops are considered low-risk samples. The distribution of the escort authorization security index is statistically analyzed, and the quantile with the highest discriminative power is selected as the threshold. When historical samples are lacking, the default threshold configured by experts can be used first and calibrated according to a preset cycle during operation. When the authorization strategy level is the ordinary authorization level or the escort gate level, the temporary access pass will be issued to the executor's terminal, and the authorization parameters of the minimum risk authority gate set and the temporary access pass will be issued to the access control equipment to realize real-time verification at the gate. When the authorization strategy level is the escort gate level, the escort gate constraints will be sent to the security escort personnel's terminal to trigger the arrival prompt and key gate confirmation process. When the authorization strategy level is the manual approval level, the explanatory materials including the domain diffusion risk index, the escort reachability credibility index, candidate path comparison and gate risk details will be included with the work order and put into the manual approval queue. For example, when a work order requires maintenance of target equipment assets in a third-floor computer room from 10:00 PM to 10:40 PM on the same day, the domain diffusion risk index increases due to the increased sensitive risk weight during the nighttime period. At the same time, the expected arrival time interval of the security escort personnel at the building lobby access control and the escort time window interval are highly matched with the service time field, resulting in a high escort reachability credibility index. The fused escort authorization security index falls within the escort gate level threshold range. A minimum risk permission door set is generated, including the building lobby access control and the computer room access control, and high-risk detour points are eliminated. The authorization parameters are set to expire from 10:00 PM to 10:50 PM. Further, an escort gate constraint condition is generated, requiring the security escort personnel to arrive at the computer room access control first and complete the arrival confirmation within 30 seconds before the execution personnel pass. Subsequently, a temporary access pass is issued to the execution personnel's terminal, and the door point authorization parameters are written into the access control device. At the same time, the escort gate constraint condition is sent to the security escort personnel's terminal, thereby achieving minimum permission and controllable escort while meeting the work order timeliness requirements.

[0031] Step five involves monitoring and controlling access control events and work order status change events during work order execution, and executing corresponding operations based on the authorization policy level. The aforementioned scheduling and allocation results, minimum risk access gate set, authorization parameters for temporary access credentials, constraints of the escort gate, and authorization policy level are implemented in real-time monitoring and control during the work order execution phase, thereby achieving a closed-loop system encompassing authorization, execution, verification, and status linkage. By monitoring and controlling access control events and work order status change events, the system can continuously perceive access control and work order progress during execution. By executing corresponding operations based on the authorization policy level, the system can take corresponding actions under different constraints of ordinary authorization level, escort gate level, and manual approval level, ensuring that work order closed-loop management goes beyond pre-authorization and allocation, covering continuous control and consistency assurance throughout the execution process.

[0032] In one specific implementation, access control events reported by the access control device and work order status change events generated during work order processing are continuously received. The two types of events are then timestamped, their sources are verified, and their fields are normalized. This ensures that access control events include at least the door identifier, passage direction, credential identifier, passage count, and event occurrence time, and that work order status change events include at least the work order identifier, status type, change time, and change reason. Based on the work order identifier, the event is then linked to the executor's identity identifier, temporary access credential identifier, minimum risk access set, and authorization policy level. The system binds access control events and serializes them according to the door point order of the access path set to form an event timeline that can be used for real-time decision-making. When a work order status change event is received, such as start execution, pause, transfer, completion, or cancellation, the validity of the temporary access certificate is synchronously controlled according to the status type. This includes activating the corresponding authorization parameters when execution starts, freezing or reissuing the authorization parameters when paused or transferred, and triggering failure handling and updating the verification basis on the access control device side when completed or cancelled. This ensures that the judgment of access control events is always consistent with the work order status. For each received access control event, real-time verification is performed to ensure it meets the authorization parameters of the minimum risk access set and temporary access pass. This verification includes at least checking whether the access point belongs to the minimum risk access set, whether the passage direction matches, whether the event occurrence time falls within the effective and expiration times, whether the number of passages exceeds the limit, and whether the pass identifier is consistent with the personnel's identity identifier. When the authorization policy level is the escort gate level, the escort gate constraints are further verified. This verification includes at least checking whether the security escort personnel's arrival confirmation has been completed at the key access point, whether the double-person arrival confirmation is satisfied, whether the passage sequence consistency is consistent with the event timeline, and whether the time coupling parameter meets the limited interval between the security escort personnel's confirmation and the personnel's pass swiping. When the verification passes, an access control command is issued to the access control device, and the release basis is written into the event timeline. Simultaneously, successful access and remaining passage count information are sent back to the personnel's terminal. Furthermore, at the escort gate level, a closed-loop notification indicating that key access point escort confirmation has been completed is simultaneously sent back to the security escort personnel's terminal. When the verification fails, a rejection command is issued to the access control device and the reason for rejection is recorded. Differentiated handling is triggered according to the authorization policy level. Under the ordinary authorization level, alarm prompts and abnormal traces are executed. Under the escort gate level, additional reminders and secondary confirmation requests for the arrival of security escort personnel are added. In the event of continuous abnormalities or increased risk, the work order status is changed to pending manual review and the evidence chain is transferred to the manual approval queue along with the work order. For example, when a work order is at the escort gate level and the minimum risk access set only includes the building lobby entrance and the computer room entrance, and the temporary access pass authorization parameters are from 22:00 to 22:50 on the same day and only two passes in the direction of entry are allowed, if the executor attempts to pass through an unauthorized entrance at 21:58, it is determined that the entrance does not belong to the minimum risk access set and the pass is rejected, while the executor is prompted to pass through according to the path; if the executor passes through the building lobby entrance at 22:06 and meets the time and number of passes restrictions, the pass is allowed and the remaining number of passes is updated; if the executor passes through the computer room entrance at 22:18 but the security escort has not yet completed the arrival confirmation or the confirmation exceeds the time coupling parameter limit interval, the pass is rejected and an arrival reminder is issued to the security escort's terminal until the security escort completes the key entrance confirmation before the executor is allowed to pass. This achieves synchronous monitoring and controllable handling of access control events and work order status change events during the work order execution process.

[0033] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. A method for AI-powered interactive customer service and closed-loop management of work orders in property management scenarios, characterized in that: include: Step 1: Obtain service request information through the interactive portal for property service recipients; Step 2: Generate structured work order data based on the service request information obtained in Step 1; Work orders are generated based on structured work order data, and then associated with the target service location and target equipment assets. Step 3: Construct a property knowledge graph based on property operation data; Based on the target service location and target equipment assets associated with the work order, a set of candidate access gates is determined, and access influence domain subgraphs and access path sets are extracted from the property knowledge graph; multi-dimensional risk weights are marked on the access influence domain subgraphs to obtain the access domain diffusion risk index; Having security escorts can increase credibility. Step 4: Generate scheduling and allocation results based on the structured work order data and the property knowledge graph; Generate a minimum risk access gate set and determine the authorization parameters for the temporary access credential; The escort authorization security index is obtained by fusing the domain diffusion risk index and the escort reachability credibility index. The escort authorization security index is compared with the preset escort threshold set to determine the authorization strategy level, and the accompanying gate constraint conditions are generated based on the authorization strategy level or the work order is transferred to the manual approval queue. Step 5: During the execution of the work order, monitor and control access control events and work order status change events, and perform corresponding operations according to the authorization policy level.

2. The AI-powered interactive customer service and work order closed-loop management method for property scenarios according to claim 1, characterized in that, The property knowledge graph includes at least spatial nodes, access gate nodes, equipment asset nodes, executor nodes, and security escort personnel nodes, and contains edges representing spatial inclusion relationships, spatial connectivity relationships, access gate connection relationships, equipment asset ownership relationships, and personnel service capabilities relationships.

3. The AI-powered interactive customer service and work order closed-loop management method for property scenarios according to claim 2, characterized in that, Multidimensional risk weights include at least spatial type risk weights, resident density risk weights, critical facility risk weights, access point security vulnerability risk weights, and time period sensitivity risk weights.

4. The AI-powered interactive customer service and work order closed-loop management method for property scenarios according to claim 1, characterized in that, The traffic influence domain subgraph contains at least a set of candidate access gates, a set of spatial nodes connected to the set of candidate access gates, a public entrance spatial node, a spatial node where the target equipment asset is located, and multiple traffic paths from the public entrance spatial node to the spatial node where the target equipment asset is located.

5. The AI-powered interactive customer service and work order closed-loop management method for property scenarios according to claim 4, characterized in that, The method for obtaining the domain diffusion risk index is as follows: Based on the multidimensional risk weights and the set of travel paths, the node risk propagation matrix is ​​obtained by calculating the risk propagation of the node risk in the subgraph of the travel influence domain. The node risk propagation matrix is ​​then normalized and weighted to obtain the domain diffusion risk index.

6. The AI-powered interactive customer service and work order closed-loop management method for property scenarios according to claim 1, characterized in that, The method for obtaining the escort reachability credibility index is as follows: generate the expected arrival time interval for each candidate access point in the candidate access point set, and generate the arrival confidence; determine the escortable time window interval; match the expected arrival time interval with the escortable time window interval and the service time field in the work order structured data to calculate the time window satisfaction; perform weighted fusion on the time window satisfaction and the arrival confidence to obtain the escort reachability credibility index.

7. The AI-powered interactive customer service and work order closed-loop management method for property scenarios according to claim 6, characterized in that, The expected arrival time interval and arrival confidence are generated as follows: Based on the real-time location data, historical trajectory data and gate congestion feature data of the security escort personnel, the expected arrival time interval of the security escort personnel to each candidate gate in the candidate gate set is calculated, and the arrival confidence is generated.

8. The AI-powered interactive customer service and work order closed-loop management method for property scenarios according to any one of claims 1-7, characterized in that, The authorization strategy levels include at least the ordinary authorization level, the escort gate level, and the manual approval level. When the authorization strategy level is the escort gate level, escort gate constraints are generated. When the authorization strategy level is the manual approval level, the work order is transferred to the manual approval queue.

9. The AI-powered interactive customer service and work order closed-loop management method for property scenarios according to claim 8, characterized in that, When the authorization policy level is the normal authorization level or the escort gate level, the temporary access pass will be issued to the executor's terminal, and the authorization parameters of the minimum risk authority gate set and the temporary access pass will be issued to the access control device; when the authorization policy level is the escort gate level, the escort gate constraints will be issued to the security escort personnel's terminal.

10. The AI-powered interactive customer service and work order closed-loop management method for property scenarios according to claim 9, characterized in that, During work order execution, for the received access control events and work order status change events, the access control events are checked in real time to see if they meet the authorization parameters of the minimum risk permission set and temporary access certificate, and the accompanying gate constraints are checked when the authorization policy level is the accompanying gate level.

Images