A hotel front and back end collaborative customer service information circulation management method and system

Abstract

The application provides a hotel front and back end cooperative customer service information circulation management method and system, which is applied to the field of information data processing; through a multi-channel information access and classification mechanism based on a hotel terminal preset, unified access and standardized processing of customer multi-source service requests are realized, and the problems of data repeated input and errors caused by different information formats are effectively solved. Through judging the identity of the service request, repeated processing is avoided, the data consistency and processing efficiency are improved, the request source information is converted into a service mirror, and combined with the preselected service branch of the customer service team, an autonomous unit is generated in real time, so that the service process is more accurate and controllable.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of information data processing, in particular to a hotel front-end and back-end collaborative customer service information circulation management method and system. BACKGROUND

[0002] In modern hotel management, customer service experience has become a key factor in measuring hotel operation efficiency and brand competitiveness. Traditional hotel customer service is mainly responsible for reception, registration and communication with customers by the front desk, while the back-end service departments such as maintenance, cleaning, guest rooms and security are responsible for specific service execution.

[0003] However, customer service requests may come from multiple channels such as telephone, front desk oral, mobile App, WeChat applet, etc. The existing terminal system is difficult to uniformly access and process, and the information formats are different, which leads to repeated entry or errors when data is commonly accessed to the terminal system. SUMMARY

[0004] The present application aims to solve the problem that customer service requests may come from multiple channels, the existing terminal system is difficult to uniformly access and process, and the information formats are different, which leads to repeated entry or errors when data is commonly accessed to the terminal system, and provides a hotel front-end and back-end collaborative customer service information circulation management method and system.

[0005] The present application solves the technical problem by using the following technical means:

[0006] The present application provides a hotel front-end and back-end collaborative customer service information circulation management method, comprising:

[0007] Based on the information access channel preset by the hotel terminal, classify the service requests uploaded by the hotel customers to the hotel terminal;

[0008] Determine whether the service request belongs to the same service content;

[0009] If yes, identify the source information of the service request, convert the source information into corresponding service mirror information, push the service mirror information to the customer service team preset by the hotel terminal, and according to the service branch preselected by the customer service team for the service mirror information, real-time generate the autonomous unit of the service request to the hotel terminal, wherein the autonomous unit specifically includes event ID, location attribute, request type and processing state;

[0010] Determine whether the autonomous unit can match the execution unit preset by the hotel terminal, wherein the execution unit specifically includes manpower type and equipment type;

[0011] If not, the information recognition engine preset by the hotel terminal is activated, a fuzzy request corresponding to the autonomous unit is constructed through the information recognition engine, the service flow of the execution unit is divided according to the fuzzy request, a task congestion heat map of the service request is generated based on the cooperative density and time lag of the service flow, and the execution resources of the hotel terminal are dynamically scheduled according to the task congestion heat map, wherein the service flow specifically includes a request intention structure, a service task disintegration chain and an execution path planning, and the task congestion heat map specifically includes a state atlas and a cooperative heat map.

[0012] Further, in the step of identifying the source information of the service request, converting the source information into corresponding service mirror information, and pushing the service mirror information to a customer service team preset by the hotel terminal, the step further includes:

[0013] Based on the content structure of the source information, a clone information object of the service request is constructed, wherein the content structure specifically includes customer identity information, request original text, trigger time and request source, and the clone information object specifically includes a customer communication clone, an execution clone, a tracking clone, a management clone and a multi-language channel adaptation clone.

[0014] It is judged whether the states of the clone information objects are uniform.

[0015] If yes, a management mechanism of the clone information object is activated, a processing period of the clone information object is allocated according to the management mechanism, and a view aggregator of the clone information object is dynamically generated, wherein the management mechanism specifically includes a read-only field, a difference update, an authority control and a version tracking, and the processing period specifically includes a generating, an already allocated, a processing, a to-be-confirmed and a closed.

[0016] Further, after the step of generating the autonomous unit of the service request to the hotel terminal according to the service branch preselected by the customer service team for the service mirror information, the step further includes:

[0017] Based on the task cooperation type of the autonomous unit, a multi-path cooperative distribution instruction of the customer service team is generated, wherein the task cooperation type specifically includes engineering maintenance, cleaning service and guest room service.

[0018] It is judged whether the multi-path cooperative distribution instruction can be allocated to idle personnel of the customer service team.

[0019] If not, the multi-path cooperative distribution instruction is listed to a preset delay processing queue, a service path of the multi-path cooperative distribution instruction is dynamically adjusted according to an entry timestamp of the delay processing queue, and a task time window of the service request is adaptively adjusted according to the service path, wherein the service path specifically includes a single-path task, a single-path task followed by subsequent dispatch, and a multi-path cooperative task.

[0020] Further, the hotel terminal is activated to preset an information recognition engine, and the step of constructing the fuzzy request corresponding to the autonomous unit by the information recognition engine further includes:

[0021] Based on a preset confidence mechanism of the information recognition engine, virtual intention data of the fuzzy request input to the information recognition engine is obtained;

[0022] It is judged whether the virtual intention data reaches a preset confidence threshold of the confidence mechanism;

[0023] If not, the fuzzy request is converted into a preset cold storage request, an intention clue of the cold storage request is retained within a preset period, information keywords repeatedly input to the hotel terminal by the hotel customer are collected, and the cold storage request is dynamically awakened according to the same word frequency of the information keywords and the intention clue.

[0024] Further, the step of judging whether the service request belongs to the same service content further includes:

[0025] Based on a preset space attribute of the hotel terminal, a request interval of the service request is obtained, wherein the space attribute specifically includes a room position and a floor position;

[0026] It is judged whether the request interval is within a preset time window;

[0027] If yes, a trigger frequency of the service request is identified, a group similarity behavior corresponding to the space attribute is obtained according to the trigger frequency, and a preset prompt content is put in a public area of the hotel terminal according to the group similarity behavior, wherein the prompt content is specifically an operation suggestion and a service mode recommendation.

[0028] Further, the step of judging whether the autonomous unit can match the preset execution unit of the hotel terminal further includes:

[0029] Based on a backhaul processing state of the autonomous unit, a processing progress of the service request is constructed, wherein the backhaul processing state specifically includes an order received, in transit, completed, and customer confirmation;

[0030] It is judged whether the processing progress is stagnant;

[0031] If yes, the area resource density of the autonomous unit is acquired, the service intensity of each area of the hotel is dynamically adjusted according to the area resource density, and the service cycle of the service request is adaptively adjusted according to the service intensity.

[0032] Further, before the step of classifying the service request uploaded by the hotel customer to the hotel terminal based on the information access channel preset by the hotel terminal, the method further comprises:

[0033] identifying the identity label preset by the hotel customer;

[0034] judging whether the identity label belongs to a preset check-in identity;

[0035] If no, the service level of the hotel customer at the hotel terminal is limited based on the identity label, and the available service content of the hotel customer is classified according to the service level.

[0036] The application also provides a hotel front-end and back-end collaborative customer service information circulation management system, comprising:

[0037] a classification module configured to classify service requests uploaded by hotel customers to the hotel terminal based on an information access channel preset by the hotel terminal;

[0038] a judgment module configured to judge whether the service requests belong to the same service content;

[0039] an execution module configured to, if yes, identify source information of the service request, convert the source information into corresponding service mirror information, push the service mirror information to a customer service team preset by the hotel terminal, and generate an autonomous unit of the service request in real time to the hotel terminal according to a service branch preselected by the customer service team for the service mirror information, wherein the autonomous unit specifically comprises an event ID, a location attribute, a request type and a processing state;

[0040] a second judgment module configured to judge whether the autonomous unit can match an execution unit preset by the hotel terminal, wherein the execution unit specifically comprises a manpower type and a device type;

[0041] a second execution module configured to, if not, activate an information recognition engine preset by the hotel terminal, construct a fuzzy request corresponding to the autonomous unit through the information recognition engine, divide a service flow of the execution unit according to the fuzzy request, generate a task congestion heat map of the service request based on the collaborative density and time lag of the service flow, and dynamically schedule execution resources of the hotel terminal according to the task congestion heat map, wherein the service flow specifically comprises a request intention structure, a service task disassembly chain and an execution path planning, and the task congestion heat map specifically comprises a state atlas and a collaborative heat map.

[0042] Further, the execution module further comprises:

[0043] A construction submodule is configured to construct the clone information object of the service request based on a content structure of the source information, wherein the content structure specifically includes customer identity information, request text, trigger time and request source, and the clone information object specifically includes a customer communication clone, an execution clone, a tracking clone, a management clone and a multilingual channel adaptation clone;

[0044] A judgment submodule is configured to judge whether the states of the clone information objects are uniform.

[0045] An execution submodule is configured to activate a management mechanism of the clone information object if the states of the clone information objects are uniform, and to allocate a processing period of the clone information object and dynamically generate a view aggregator of the clone information object according to the management mechanism, wherein the management mechanism specifically includes a read-only field, a difference update, an authority control and a version tracking, and the processing period specifically includes a generating, an already allocated, a processing, a to-be-confirmed and a closed.

[0046] Further, the execution module further comprises:

[0047] A generation module is configured to generate a multi-path cooperative distribution instruction of the customer service team based on a task cooperation type of the autonomous unit, wherein the task cooperation type specifically includes engineering maintenance, cleaning service and room service.

[0048] A third judgment module is configured to judge whether the multi-path cooperative distribution instruction can be allocated to idle personnel of the customer service team.

[0049] A third execution module is configured to list the multi-path cooperative distribution instruction to a preset delay processing queue if the multi-path cooperative distribution instruction cannot be allocated to the idle personnel of the customer service team, dynamically adjust a service path of the multi-path cooperative distribution instruction according to an entry time stamp of the delay processing queue, and adaptively adjust a task time window of the service request according to the service path, wherein the service path specifically includes a single-path task, a single-path task and subsequent supplementary allocation and a multi-path cooperative task.

[0050] The present application provides a hotel front-end and back-end cooperative customer service information circulation management method and system, which has the following advantages:

[0051] The application realizes unified access and standardized processing of customer multi-source service requests through a multi-channel information access and classification mechanism preset based on a hotel terminal, effectively solves the problems of repeated data entry and errors caused by different information formats. By judging the identity of the service request, repeated processing is avoided, the data consistency and processing efficiency are improved, the request source information is converted into a service mirror, and combined with the preselected service branch of the customer service team, an autonomous unit is generated in real time, so that the service process is more accurate and controllable. BRIEF DESCRIPTION OF DRAWINGS

[0052] Figure 1 A flowchart of an embodiment of the hotel front-end and back-end collaborative customer service information circulation management method of the application;

[0053] Figure 2 A structural block diagram of an embodiment of the hotel front-end and back-end collaborative customer service information circulation management system of the application. DETAILED DESCRIPTION

[0054] It should be understood that the specific embodiments described herein are only used to explain the application and do not limit the application. The purpose, functional features and advantages of the application will be further described with reference to the embodiments and the accompanying drawings.

[0055] The technical solutions of the embodiments of the application will be described clearly and completely below with reference to the accompanying drawings of the embodiments of the application. Obviously, the described embodiments are only part of the embodiments of the application, not all the embodiments. Based on the embodiments in the application, all other embodiments obtained by those skilled in the art without creative labor are within the scope of protection of the application.

[0056] Reference is made to the accompanying drawings Figure 1 The hotel front-end and back-end collaborative customer service information circulation management method in an embodiment of the application comprises:

[0057] S1: Based on the information access channel preset by the hotel terminal, classify the service requests uploaded by the hotel customers to the hotel terminal;

[0058] S2: Determine whether the service request belongs to the same service content;

[0059] S3: If yes, identify the source information of the service request, convert the source information into corresponding service mirror information, push the service mirror information to the customer service team preset by the hotel terminal, and according to the service branch preselected by the customer service team for the service mirror information, generate an autonomous unit of the service request to the hotel terminal in real time, wherein the autonomous unit specifically includes event ID, location attribute, request type and processing state;

[0060] S4: judging whether the autonomous unit can match the execution unit preset by the hotel terminal, wherein the execution unit specifically includes manpower type and equipment type;

[0061] S5: if not, activating an information recognition engine preset by the hotel terminal, constructing a fuzzy request corresponding to the autonomous unit through the information recognition engine, dividing a service flow of the execution unit according to the fuzzy request, generating a task congestion heat map of the service request based on the service flow of the service flow and time lag, and dynamically scheduling the execution resource of the hotel terminal according to the task congestion heat map, wherein the service flow specifically includes a request intention structure, a service task disassembly chain and an execution path planning, and the task congestion heat map specifically includes a state atlas and a cooperation heat map.

[0062] In the present embodiment, the system accesses the information channel pre-set by the hotel terminal based on the information pre-set by the hotel terminal, classifies the service requests uploaded to the hotel terminal by the hotel customers, and then judges whether the service requests belong to the same service content to perform corresponding steps; for example, when the system judges that the service requests uploaded to the hotel terminal by the hotel customers do not belong to the same service content, the system considers that the multiple service requests currently received have obvious differences in intention, execution mode, resource demand or processing logic, different customers or different rooms submit requests with similar content but different targets, the system constructs an independent service mirror information structure for each request with different service content, keeps the logic isolated between requests, avoids confusion, each mirror binds its source, timestamp, context data, at the same time, different service requests are distributed to professional customer service groups to avoid resource conflicts or repeated responses caused by “one order multiple investment”, and if multiple different requests come from the same customer (or room) and have no conflict, they can be guided to be combined into “parallel service flow” to improve efficiency, otherwise, they are scheduled in turn according to priority; for example, when the system judges that the service requests uploaded to the hotel terminal by the hotel customers belong to the same service content, the system considers that the multiple service requests currently received may have repeated service events, the system identifies the source information of the service requests, converts the source information into corresponding service mirror information, pushes the service mirror information to the pre-selected service branches of the service mirror information in the pre-selected customer service team of the hotel terminal, and generates an autonomous unit of the service request to the hotel terminal in real time, the autonomous unit specifically includes event ID, location attribute, request type and processing state; the system identifies the source information of multiple requests, combines multiple requests belonging to the same service content to generate an autonomous unit, avoids multiple repeated work orders entering the scheduling queue, this deduplication mechanism can prevent human resources and equipment resources from being repeatedly called, especially during peak hours or continuous customer operations (such as repeatedly pressing the service button), which can significantly reduce system congestion rate and execution interference, improve the coordination and execution efficiency of the overall service, at the same time, the original request is converted into standardized service mirror information, which can unify the expression methods of different sources (such as App, voice, front desk), avoid the understanding deviation of the service intention by the customer service or the system, the service mirror contains unified field structure, such as request type, location attribute, timestamp, etc., which lays the information foundation for subsequent generation of autonomous unit, ensures that the information received at the execution level is clear, complete and operable, and different customer service teams can pre-set the processing branches (such as delivery, maintenance, cleaning, etc.) of the corresponding service mirror, the system automatically matches the optimal service path according to the mirror content to realize intelligent classification and accurate distribution of service requests, the generated autonomous unit not only has a clear event ID and state, but also can be quickly routed to the corresponding post or equipment execution node, thereby realizing highly automated and fine management of service process;The system then determines whether the autonomous unit can match the execution unit preset by the hotel terminal. The execution unit specifically includes manpower and equipment to execute the corresponding steps. For example, when the system determines that the autonomous unit of the service request can match the execution unit preset by the hotel terminal, the system considers that the current task can be identified and the corresponding execution entity can be found. The system assigns the task to the mobile terminal of the corresponding post personnel (such as the PDA of the cleaner or the engineering maintenance tablet), including detailed task content, location, and response time limit, triggers device instructions (such as intelligent delivery robot scheduling, elevator permission opening, air conditioning adjustment system automatic response, etc.), executes specific operations, updates the task status from “to be processed” to “in execution” or “has been assigned”, binds the identification of the execution unit (such as personnel ID, device number), ensures that the subsequent tracking is clear and controllable, and continuously listens to the execution feedback of the execution unit, collects the execution progress, completion status, and abnormal alarm at regular intervals, and feeds back to the hotel central control system or customer service terminal. If necessary, it can support automatic follow-up or abnormal manual processing. For example, when the system determines that the autonomous unit of the service request cannot match the execution unit preset by the hotel terminal, the system considers that the current task can be identified, but the corresponding execution entity cannot be found at present. The system activates the information recognition engine preset by the hotel terminal. Through the information recognition engine, the fuzzy request corresponding to the autonomous unit is constructed. According to the fuzzy request, the service flow of the execution unit is divided, which specifically includes the request intention structure, the service task decomposition chain, and the execution path planning. Based on the coordination density and time lag of the service flow, the task congestion heat map of the service request is generated, which specifically includes the state atlas and the coordination heat map. The system identifies the potential intention through the information recognition engine, constructs the fuzzy request, avoids the interruption problem of “request recognition without execution path”, and can further intelligently generate the execution scheme through intention understanding and task decomposition, filling the short board of traditional systems in handling non-standard tasks. At the same time, by constructing the “service flow” (including the request intention structure, the task decomposition chain, and the execution path planning), the system can refine the fuzzy or composite request into executable standard sub-tasks, plan and schedule the path according to the priority and accessibility. This way makes complex tasks can be decomposed into multiple concurrent processing units, improves execution efficiency and reduces single-point bottlenecks. In combination with the coordination density and time lag of the service flow, the task congestion heat map (including the state atlas and the coordination heat map) is generated, which directly shows the task distribution and resource pressure. This can assist the system in executing flow limiting, detour scheduling, or task rearrangement in high-load areas to avoid resource overload in some areas while other areas are idle, and ultimately realize the optimization of resource allocation and the balance of service response.

[0063] It should be noted that the source information of the service request is identified, the source information is converted into corresponding service image information, the service image information is pushed to the customer service team preset by the hotel terminal, and based on the service branch pre-selected by the customer service team for the service image information, the autonomous unit of the service request is generated in real time to the hotel terminal. The specific example is as follows:

[0064] Assume that guest Mr. Zhang is staying in room 1802 of the hotel. He sends the same service request at 10:06 in three different ways:

[0065] Click the "Add a Towel" button through the hotel app;

[0066] Call room service and say, "Please send up a towel";

[0067] The voice dialogue speaker said "Can you send me another white bath towel?";

[0068] The system processing process is detailed as follows:

[0069] Step 1: Identify the source information of the service request. The system receives service requests from different channels and records the following source information:

[0070] Channel Source information field APP Customer ID: ZH1802, timestamp: 10:06:12, room number: 1802, request code: REQ001 Phone Call ID: #1802TEL, transcribed text: "Please send a bath towel", room number identification: 1802 Voice assistant Device ID: VOICE-1802, voice text: "Can you send another white bath towel", room number: 1802

[0071] The system will identify these requests as coming from the same source (same customer, same room number, close in time) and proceed to the next step;

[0072] Step 2: Convert to service image information. The system standardizes different expressions into a structured "service image information":

[0073] {

[0074] "room_number":"1802",

[0075] "guest_id":"ZH1802",

[0076] "intent":"Add extra items",

[0077] "item":"Bath towel",

[0078] "item_color":"white",

[0079] "channel":["APP","Phone","Voice Assistant"],

[0080] "timestamp_range":["10:06:12","10:06:45"],

[0081] "priority":"ordinary",

[0082] "service_id":"SRV-MIRROR-001"

[0083] }

[0084] This step completes the process of unifying "unstructured voice, graphic click, and semantically ambiguous" requests into "service image information";

[0085] Step 3: Push the request to the customer service team. The system determines that the request is "room extra items" based on the image content and automatically pushes it to the customer service team of the "room service group". This team system has a preset response branch strategy. For example, if the request is "bath towel", the delivery robot will be given priority. If the robot is busy, the floor attendant will be dispatched for manual delivery.

[0086] Step 4: Generate an autonomous unit. After receiving the image information, the customer service system automatically generates a schedulable service autonomous unit as follows:

[0087] {

[0088] "event_id":"EVT-1802-20250610-001",

[0089] "location":{

[0090] "room":"1802",

[0091] "floor":18

[0092] },

[0093] "request_type":"Send bath towel",

[0094] "item_detail":{

[0095] "name":"Bath towel",

[0096] "color":"white",

[0097] "quantity":1

[0098] },

[0099] "channel_merged":["APP","TEL","VOICE"],

[0100] "status": "pending scheduling",

[0101] "dispatch_type":"Device Priority",

[0102] "expected_eta": "5 minutes"}

[0103] This unit is then read by the delivery robot scheduling system to execute the delivery operation; if the scheduling fails, it is transferred to the manual task pool;

[0104] To summarize, the above example avoids duplicate work by abstracting ambiguous requests from three different channels into a service mirror. At the same time, since the information source is consistent but the expression is different, the system can avoid resource waste or task conflicts caused by duplicate requests. There is no human intervention in the entire process. The system automatically responds and dispatches based on "semantic standardization + task autonomy unit", greatly improving response efficiency and customer satisfaction.

[0105] It should be supplemented that the information recognition engine preset in the hotel terminal is activated, and a fuzzy request corresponding to the autonomous unit is constructed through the information recognition engine. The service flow of the execution unit is divided according to the fuzzy request, and a task congestion heat map of the service request is generated based on the collaborative density and time lag of the service flow. The execution resources of the hotel terminal are dynamically scheduled according to the task congestion heat map. A specific example is as follows:

[0106] Suppose a customer is in room 1908 of a hotel and asks the voice assistant, "Something's not right with my room. Could you please come take a look?" Because this request is semantically ambiguous and can't be directly categorized as "delivery," "repair," or "customer complaint," the system performs a detailed breakdown of the steps:

[0107] First, the information recognition engine is activated. The system recognizes the key words "something is wrong" and "come and see" in the voice, and determines it as an unstructured request. It automatically activates the preset information recognition engine in the terminal and starts the fuzzy request construction process. By constructing the fuzzy request, the system combines the context data to generate the following "fuzzy request object":

[0108] {

[0109] "Room Number":"1908",

[0110] "Trigger Time":"2025-06-10T11:42:00",

[0111] "Keywords": ["something's wrong", "take a look"],

[0112] "Historical behavior": ["Ordered room water 2 hours ago", "Complained about air conditioning noise at night"],

[0113] "Room status data": ["Door magnetic abnormal opening and closing multiple times", "Temperature and humidity sensor short disconnection"],

[0114] "System determines intention": "Potential device anomaly / emotional pacification needs / security risk investigation"

[0115] Then the system divides the service flow, generates the service flow according to the above fuzzy request, and the process includes identifying the request intention structure as "non-standard security request", and suggesting the service type: ① room state detection → ② security personnel visit → ③ front desk follow-up inquiry, background call room sensor data and access control record, arrange engineering department personnel to do safety inspection on air conditioner and power supply, notify the front desk to call the customer to confirm whether everything is normal, arrange the floor service robot to send a access control prompt card first, arrange the nearest staff to relay to the door, reduce the waiting time;

[0116] Then generate the task congestion heat map, and the heat map is constructed as follows:

[0117] State atlas data:

[0118] Node Current task number Average response delay Distance from room 1908 Security A (1902) 1 3 minutes 1 room Engineering B (downstairs) 3 15 minutes 1 floor Service robot C Idle Instant Same floor

[0119] Collaborative heat map results:

[0120] The current floor has multiple "non-standard requests" concentrated in 1910~1906, indicating that there may be systematic failures in the area, which need to be synchronized with the engineering department;

[0121] Finally, dynamically schedule the execution resources, and the system decision is as follows:

[0122] Arrange service robot C to deliver "room safety reminder card" immediately, first pacify the customer, and staff A will visit 1908 after finishing 1902, engineer B is overloaded, and remote diagnose the air conditioner equipment state, send a memo to the front desk, and list "multiple room abnormalities on this floor" in the batch processing suggestion, and the system will feedback the customer information (such as through the voice assistant), "we have arranged personnel to confirm your room state, please wait, service robot will deliver a prompt card, if there is more specific abnormal content, you can also continue to tell me".

[0123] In the above example, even if the semantics are fuzzy, the system can preliminarily process and schedule in a structured way, identify potential batch problems in advance through collaborative heat maps, reduce repeated response costs, improve service accuracy and customer experience, and relieve the pressure on front-line personnel.

[0124] In this embodiment, the source information of the service request is identified, the source information is converted into corresponding service mirror information, and the service mirror information is pushed to the preset customer service team of the hotel terminal in step S3.

[0125] S31: constructing a clone information object of the service request based on a content structure of the source information, wherein the content structure specifically comprises customer identity information, request original text, trigger time and request source, and the clone information object specifically comprises a customer communication clone, an execution clone, a tracking clone, a management clone and a multi-language channel adaptation clone;

[0126] S32: judging whether states of each of the clone information objects are unified;

[0127] S33: if yes, activating a management mechanism of the clone information object, allocating a processing period of the clone information object according to the management mechanism, and dynamically generating a view aggregator of the clone information object, wherein the management mechanism specifically comprises a read-only field, a differential update, an authority control and a version tracking, and the processing period specifically comprises a generating, an already allocated, a processing, a to-be-confirmed and a closed.

[0128] In the embodiment, the system constructs a clone information object of the service request based on a content structure of the source information, the content structure specifically includes customer identity information, a request original text, a triggering time and a request source, the clone information object specifically includes a customer communication clone, an execution clone, a tracking clone, a management clone and a multilingual channel adaptation clone, and then the system judges whether the states of the clone information objects are uniform to perform corresponding steps; for example, when the system determines that the states of the clone information objects are not uniform, the system considers that a state difference or data asynchronization is generated in multi-link circulation of the current service request, that is, the same customer request presents inconsistent service states or interpretation results in different “clones” such as customer communication, execution layer feedback, state tracking, background management and multilingual translation adaptation, the system compares task IDs, time stamps and execution states in the customer communication clone, the execution clone and the tracking clone to calculate a consistency score; if the score is lower than a threshold value, the system automatically enters a “state reconstruction” process, reinfers an initial service intention from the “request original text + customer identity + triggering time” in the source information, and preferentially allocates a manual customer service if the triggered abnormality is high in severity (such as a safety type or a complaint type), or automatically pulls a standard response path to reissue an execution task if it is a regular service item; for example, when the system determines that the states of the clone information objects are uniform, the system considers that information between each link of the current service request from receiving to execution, tracking, feedback, management and multilingual adaptation is synchronous and consistent, and state drift, interpretation conflict or execution deviation does not occur, the system activates a management mechanism of the clone information object, the management mechanism specifically includes a read-only field, a difference update, a permission control and a version tracking, allocates a processing period of the clone information object according to the management mechanism, and the processing period specifically includes a generated state, an already assigned state, a processing state, a to-be-confirmed state and a closed state.The system can effectively prevent the service request from being tampered or mismodified in the multiple processing process by activating the management mechanism of the clone information object and introducing read-only fields, differential updates, permission control and version tracking. Even if the service request goes through multiple sub-departments and multiple operation nodes, the system can still maintain a highly consistent state view and record the operation history by version, thereby improving the transparency and credibility of the system. At the same time, the service request is divided into standardized processing cycle stages such as "in generation → assigned → in processing → to be confirmed → closed", which can make the system have clear identification and process control ability for service status. This not only facilitates the automatic scheduling of human or equipment resources, but also unifies the understanding of service status by different departments, effectively avoiding misunderstanding, repeated processing or state misplacement. Moreover, each clone object is subject to unified management mechanism, ensuring its behavior has boundary constraints (such as read-only restrictions), dynamic response (such as differential updates), access security (such as permission control) and historical traceability (such as version tracking). This helps to control the access and modification of data by different permission levels, accurately identifies abnormal updates or tampering behavior, and supports cross-language, cross-role and cross-platform service request synchronization and coordination.

[0129] In the embodiment, after the step S3 of generating the autonomous unit of the service request to the hotel terminal in real time according to the service branch preselected by the customer service team from the service mirror image information, the method further comprises:

[0130] S301: generating a multi-path collaborative distribution instruction of the customer service team based on a task collaboration type of the autonomous unit, wherein the task collaboration type specifically includes engineering maintenance, cleaning service and guest room service;

[0131] S302: judging whether the multi-path collaborative distribution instruction can be distributed to idle personnel of the customer service team;

[0132] S303: if not, listing the multi-path collaborative distribution instruction to a preset delay processing queue, dynamically adjusting a service path of the multi-path collaborative distribution instruction according to an entry time stamp of the delay processing queue, and adaptively adjusting a task time window of the service request according to the service path, wherein the service path specifically includes single task, single task followed by subsequent supplementary assignment and multi-path collaborative task.

[0133] In the embodiment, the system generates multi-path collaborative distribution instructions of the customer service team based on the task cooperation types of the autonomous units, the task cooperation types specifically including engineering maintenance, cleaning service and room service, and then the system determines whether the multi-path collaborative distribution instructions can be distributed to idle personnel of the customer service team to perform corresponding steps; for example, when the system determines that the multi-path collaborative distribution instructions of the customer service team can be distributed to the idle personnel of the customer service team, the system considers that the current human resource pool is schedulable and available, and the indexes such as skill matching degree, time window and work load between the task and the personnel meet the scheduling conditions, the system considers that the task can be received and executed immediately and efficiently, the system pushes each multi-path collaborative distribution instruction to the terminal of the matched idle personnel, marks the task as “distributed” state, generates a task tracking node in the system for subsequent state collection, feedback and time limit monitoring, at the same time, if the service request corresponds to multiple tasks (for example, a “leakage event” may involve maintenance + cleaning), the system will issue multiple distribution instructions to different team members in parallel, bind them as a whole in the background, and perform overall state normalization management, and can synchronously display “task has been assigned, and the service will be provided within a few minutes” to the customer to improve customer experience and transparency during the waiting process; for example, when the system determines that the multi-path collaborative distribution instructions of the customer service team cannot be distributed to the idle personnel of the customer service team, the system considers that the current human resource pool is not available, and the task and the personnel cannot be matched, the system lists the multi-path collaborative distribution instructions in the pre-set delay processing queue, dynamically adjusts the service path of the multi-path collaborative distribution instructions according to the time stamp of the multi-path collaborative distribution instructions entering the delay processing queue, the service path specifically includes single-path task, single-path task and subsequent supplementary distribution, and multi-path collaborative task, and the task time window of the service request is adaptively adjusted according to the service path; the system avoids blind repeated attempts and scheduling conflicts of the task when the resources are insufficient by adding the task that cannot be immediately assigned to the delay processing queue, and can process the task in order according to the emergency degree and waiting time length by combining the time stamp and the queue rule, so as to ensure that the task execution is more planned and buffered, improve the stability and fairness of the overall resource scheduling, dynamically adjust the original multi-path collaborative task to a single-path task, a supplementary distribution task or a split execution path according to the executable conditions, so that the system can still maintain uninterrupted service when the personnel are insufficient, the adaptive strategy of the task path enhances the flexibility of the task model, effectively improves the response ability of the system in complex scenarios such as peak period, sudden offline of personnel and dense service requests, and the system adaptively updates the task time window of the service request (such as postponing the estimated arrival time, supplementary state notification, etc.) according to the task type and the service path adjustment, avoids the experience loss caused by long-time no feedback of the customer, provides more reasonable buffer time for subsequent execution, maintains the right to know and expectation management of the customer on the task state by cooperating with the customer front-end state synchronization, and improves the overall satisfaction.

[0134] It should be noted that the multi-path collaborative distribution instruction is listed to the preset delay processing queue, the service path of the multi-path collaborative distribution instruction is dynamically adjusted according to the entry time stamp of the delay processing queue, and the task time window of the service request is adaptively adjusted according to the service path, and specific examples are as follows:

[0135] Suppose that in a high-end hotel, the guest of room 501 sends a service request through the intelligent terminal in the room: "the hand washing room has serious water accumulation, and it is suspected that the drain pipe is blocked, please handle as soon as possible";;

[0136] The original service path (multi-path collaborative) is planned, and the system intelligently analyzes that the task needs to be completed by three types of execution units according to the keywords "water accumulation", "drain pipe" and "handling":

[0137] Engineering maintenance personnel: responsible for detecting the drainage system and dredging;

[0138] Cleaning personnel: responsible for cleaning water accumulation and wiping the floor;

[0139] Room service personnel: supplement bath towel and replace wet supplies;

[0140] The system queries the available resources as follows,

[0141] Engineering maintenance personnel: currently all occupied, and the estimated idle time is more than 25 minutes;

[0142] Cleaning personnel: busy, no clear release time;

[0143] Room service personnel: one is idle, located near room 502;

[0144] Therefore, the delay processing mechanism is started, the delay processing queue is listed, the maintenance and cleaning tasks are packaged as a "delay processing task unit", the entry time stamp (such as 10:30:14) is marked, and the priority (such as medium to high) is recorded, and the room service personnel task is executed as an instant task;

[0145] Then the service path is dynamically adjusted, and the service path variant is generated as follows,

[0146] Path A (instantly feasible): immediately assign the room service personnel near room 502 to enter room 501 first, deliver clean bath towel and replace wet floor mat, and appease the customer;

[0147] Path B (subsequently dispatched): the cleaning and maintenance tasks are set as delay tasks, and the system continuously monitors the resource availability state;

[0148] System internal remarks: "the service request is converted into a phased completion mode";

[0149] Then the task time window is updated adaptively. By adjusting the task time window, the original setting of "complete all services within 15 minutes" is adjusted as follows,

[0150] Room replenishment task: handle immediately (within 5 minutes);

[0151] Maintenance task: expected to start at the earliest before 10:55;

[0152] Cleaning task: expected to follow up within 10 minutes after maintenance is completed;

[0153] Finally, the feedback is synchronized to the customer interface (such as TV screen, App, voice broadcast) as follows,

[0154] "Hello, service request has been started, clean towels have been arranged, and drainage inspection and cleaning will be completed after personnel are available. Thank you for your understanding;

[0155] Regarding the delay queue monitoring and subsequent scheduling, the task heat will increase over time. After 10 minutes, the system finds that the task has not been completed, the queue timestamp continues to increase, and the system automatically raises the priority by one level. When a cleaning staff member's task is completed, the system immediately schedules to go to room 501. After the engineering maintenance staff member completes the previous task (about 11:00), they are automatically assigned to room 501 to handle the remaining issues;

[0156] Finally, the service completion feedback is provided. After all services are completed, the system summarizes the task execution track and time node, records the start and end times of the three stages of service, synchronizes to the management background for subsequent scheduling strategy optimization, and sends a complete service summary card (including apology, points reward, etc.) to the customer;

[0157] In summary, the above example shows that the system can provide timely response and emotional relief to customers even if it cannot allocate resources to serve them in a timely manner by reasonably arranging hotel resource allocation. The system also reasonably utilizes local resources to avoid complete denial of service, forms a service data closed loop in the back end, which helps subsequent prediction and deployment optimization, and realizes the intelligent scheduling capability of "peak shifting, segmented processing, and task reconstruction" during peak periods.

[0158] In the embodiment, the information recognition engine preset in the hotel terminal is activated, and the step S5 of constructing the fuzzy request corresponding to the autonomous unit further includes:

[0159] S51: based on the confidence mechanism preset in the information recognition engine, obtaining virtual intention data after the fuzzy request is input into the information recognition engine;

[0160] S52: determining whether the virtual intention data reaches the confidence threshold preset in the confidence mechanism;

[0161] S53: If not, the fuzzy request is converted into a preset cold storage request, the intention clue of the cold storage request is kept for a preset period, information keywords repeatedly input by the hotel customer into the hotel terminal are collected, and the cold storage request is dynamically woken up according to the same word frequency of the information keywords and the intention clue.

[0162] In the embodiment, the system obtains the virtual intent data of the fuzzy request input into the information recognition engine based on the confidence mechanism preset by the information recognition engine, and then determines whether the virtual intent data reaches the confidence threshold preset by the confidence mechanism to execute corresponding steps; for example, when the system determines that the virtual intent data of the fuzzy request input into the information recognition engine can reach the confidence threshold preset by the confidence mechanism, the system considers that the current fuzzy request can be accurately analyzed by the system to obtain the real intent of the customer, and the system obtains the core fields such as the user intent keyword, target location, and expected time limit from the analysis result, marks it as "identified virtual intent", and attaches a confidence value. At the same time, according to the identified intent, a corresponding service autonomous unit is constructed, the unit is still marked as "generated by fuzzy converted intent", but can be included in the formal scheduling system, and the virtual intent is automatically mounted to one or more existing service branches. If there are multiple candidate service branches, the one with the highest matching degree is automatically selected according to the confidence value and service semantic priority; for example, when the system determines that the virtual intent data of the fuzzy request input into the information recognition engine cannot reach the confidence threshold preset by the confidence mechanism, the system considers that the current fuzzy request cannot be analyzed by the system to obtain the real intent of the customer, and the system converts the fuzzy request into a cold storage request. The intent clues of the cold storage request are preserved within a preset period of time, and the cold storage request is dynamically awakened according to the same word frequency of the information keywords and the intent clues collected by the hotel customer repeatedly input into the hotel terminal. When the system cannot clearly identify the intent of the customer, it does not blindly generate tasks, but stores the request in cold storage to prevent resource waste, service deviation, or customer dissatisfaction caused by incorrect scheduling. Through the "intent cold storage + re-determination" mechanism, the service execution is ensured to have a high confidence as a prerequisite, and each operation is ensured to be more targeted and accurate. At the same time, the system saves the semantic clues and keyword residual data of the un-identified request, gradually constructs a language preference model of the hotel customer behavior, and when the same customer inputs similar expressions in a short period of time, the system can be awakened based on "keyword co-occurrence degree" and "context correlation degree" to make the originally un-identified request have context regeneration ability. The system continuously optimizes the understanding of natural language, and through the "cold storage request + dynamic awakening" method, the system establishes a fault-tolerant waiting mechanism. Even if the first input fails, it will not be discarded immediately, but will wait for the customer's subsequent behavior to "confirm again". This strategy not only improves the tolerance of the system to fuzzy and ambiguous semantics, but also provides a more relaxed operation tolerance for the customer, enhancing the user experience.

[0163] In the embodiment, the step S2 of determining whether the service requests belong to the same service content further includes:

[0164] S21: Obtain a request interval of the service request based on a space attribute preset by the hotel terminal, wherein the space attribute specifically comprises a room position and a floor position;

[0165] S22: Determine whether the request interval is within a preset time window.

[0166] S23: If yes, identify a trigger frequency of the service request, obtain a group similarity behavior corresponding to the space attribute according to the trigger frequency, and deliver preset prompt content in a public area preset by the hotel terminal according to the group similarity behavior, wherein the prompt content specifically comprises operation suggestions and service mode recommendations.

[0167] In the embodiment, the system obtains the request intervals of the service requests based on the space attributes of the hotel terminal, and the space attributes specifically include room position and floor position. Then, the system determines whether the request intervals are within the pre-set time window to execute corresponding steps. For example, when the system determines that the request intervals of the service requests are not within the pre-set time window, the system considers that the current service requests present non-short-time concentration characteristics in the time dimension, and there is no strong correlation or close trigger logic between the requests. The system considers that the requests have spatial correlation but lack time coincidence, and do not constitute the processing conditions of “task merging” or “batch triggering”. Therefore, the requests should be processed independently according to the respective event IDs, and autonomous units are generated respectively to avoid merging unrelated events into the same service flow, prevent misaggregation, and update the service state nodes of the corresponding rooms or floors in the hotel terminal space attribute graph. The system marks that there are “scattered requests” distributed in the region. If these nodes continue to heat up in a short period of time, the system can set a threshold to trigger regional inspection or preventive task dispatching (such as notifying the cleaning staff and equipment self-checking robots). If the same space region repeatedly appears in the space-time and the content of the service requests is similar (such as multiple customers reporting that the TV cannot be used at different times), the system considers that there is a risk of repeated complaints or an abnormal trend of equipment, and can enter the “abnormal service request aggregation model” to prompt the management personnel to intervene and check. For example, when the system determines that the request intervals of the service requests are within the pre-set time window, the system considers that the requests have obvious concentration or continuity in the time dimension, belong to repeated submission of the same customer, or are prompt requests before the task is completed. The system identifies the trigger frequency of the service requests, obtains the group similarity behavior corresponding to the space attributes according to different trigger frequencies, and puts pre-set prompt contents in the pre-set public areas of the hotel terminal according to the group similarity behavior. The prompt contents specifically include operation suggestions and service mode recommendations.The system automatically determines possible repeated submission or task prompting behavior by identifying the concentration and high frequency of the request, thereby avoiding multiple order placement and repeated processing for the same service event, optimizing resource scheduling efficiency, reducing human waste and device calling redundancy, and improving the intelligent level of the overall service process. At the same time, by analyzing the similarity of group behavior, personalized and perceptible prompt information (such as "air conditioner restart needs to wait for 3 minutes, please do not repeat submission" and "if you need to speed up the processing, please confirm the service status through the guest room tablet") is provided in public areas, which can guide the service rhythm without interfering with the user's privacy, relieve the anxiety of customers caused by waiting or misoperation, improve service transparency and user experience, and through group behavior analysis (for example, multiple adjacent rooms submit "room temperature abnormality" at the same time), the system can determine potential device or systematic problems, provide public prompt content in time, trigger background diagnosis tasks, realize proactive intervention and prevention before the problem reaches the peak, and enhance the robustness and self-healing ability of the hotel service system.

[0168] In the embodiment, the step S4 of judging whether the autonomous unit can match the execution unit preset by the hotel terminal further includes:

[0169] S41: based on the backhaul processing state of the autonomous unit, a processing progress of the service request is constructed, wherein the backhaul processing state specifically includes accepted order, in transit, completed and customer confirmation;

[0170] S42: judging whether the processing progress is stagnant;

[0171] S43: if yes, the regional resource density of the autonomous unit is obtained, the service intensity of each region of the hotel is dynamically adjusted according to the regional resource density, and the service cycle of the service request is adaptively adjusted according to the service intensity.

[0172] In the embodiment, the system constructs the processing progress of the service request based on the backhaul processing state of the autonomous unit, and the backhaul processing state specifically includes order accepted, in transit, completed, and customer confirmation. Then, the system determines whether the processing progress is stagnant to perform corresponding steps. For example, when the system determines that the processing progress of the service request is not stagnant, the system considers that the current service request is in a normal flow state from dispatch to execution process, and each link of the execution unit is feeding back the processing state according to the process, without abnormal delay or interruption. The system updates the processing record of the service request in real time according to the state of the execution unit (human or device) returned at regular intervals, ensures the traceability of the entire service link, synchronizes the current processing progress to the customer device (such as a room tablet, WeChat applet, etc.) in real time, and prompts the customer to the current state through the form of a "service progress bar" or a "state reminder card", thereby enhancing the perceptibility and satisfaction of the customer during waiting. Moreover, since the processing progress is in a coherent and smooth state, the system does not need to enable an abnormal processing mechanism (such as task reminder, heat map congestion analysis, or replacement mechanism), maintains the automatic execution mode of the current service chain, effectively reduces the management intervention and scheduling resource load, and so on. For example, when the system determines that the processing progress of the service request is stagnant, the system considers that the current service request is in an abnormal flow state. The system acquires the regional resource density of the autonomous unit, dynamically adjusts the service intensity of each region of the hotel according to different regional resource densities, and adaptively adjusts the service period of the service request according to the service intensity. The system can quickly identify the "service bottleneck region" or "human resource shortage area" by analyzing the resource density of the region where the autonomous unit is located (such as the number of execution units in the same region, the availability of human resources, the response frequency of equipment, etc.), avoid blindly global scheduling, reduce the waste of scheduling resources, realize the precise intervention of the abnormal flow request, and so on. According to the service intensity (i.e., the ratio of the aggregation degree of service requests in a unit time to the task processing capacity), the system can set a more reasonable service time window or appropriately delay the response period for new service requests in the region, so that the processing link of the high-density region obtains a breathing opportunity, guarantees the priority of the existing task closed loop, avoids the influence of long-time suspended tasks on the overall service quality, and so on. Moreover, the adaptive adjustment of the service period not only can improve the sustainability of service execution in the congested area, but also can reduce the customer complaint rate caused by task accumulation, enhance the response flexibility of the system to sudden concentrated service requests, and so on. The system controls the processing rhythm of the task in a fine-grained manner, so that the overall hotel service network remains in a "flexible and controllable" state.

[0173] It should be noted that the regional resource density of the autonomous unit is acquired, the service intensity of each region of the hotel is dynamically adjusted according to the regional resource density, the service period of the service request is adaptively adjusted according to the service intensity, and specific examples are as follows:

[0174] Suppose a large chain hotel welcomes a conference group on a weekend night, and the 16th and 17th floors are assigned to the group members; among them, the east side guest room area of the 16th floor has a concentrated outbreak of service requests in a short period of time, and the system records as follows:

[0175] Room number Service request Submission time 1602 Replace towel 18:03 1605 Air conditioner not cooling (repair) 18:04 1607 Send bottled water 18:05 1609 Wall plug power not working (repair) 18:07 1603 Send toothbrush 18:08 1608 Carpet has stains (cleaning request) 18:09 1604 Air conditioner still not working (repeat prompting) 18:10

[0176] At 18:11, the system automatically analyzes the task situation of the east side of the 16th floor, and the process is as follows:

[0177] Step 1, get the area resource density,

[0178] Current total number of service requests: 7

[0179] Idle manpower resources: 1 floor attendant (No. E01)

[0180] Idle equipment resources: 1 service robot (No. R02)

[0181] Area resource density calculation: total number of available resources = 2 (E01 + R02), total number of requests = 7, → resource density ≈ 0.29, below the "resource shortage" threshold;

[0182] Step 2, dynamically adjust the service intensity, the system judges that there is a problem of high service load on the east side of the 16th floor, and immediately executes,

[0183] Service task reorganization: combine "1603 send toothbrush" and "1607 send water" into a unified delivery task, execute "1602 change towels" by robot (bag type delivery), prioritize "1605 air conditioner not cooling" for engineering department personnel remote initial diagnosis, check the status of service attendant E02 on the 17th floor, if idle, then allocate "1609 wall plug power no power" task to E02 to handle, at the same time, eliminate duplicate / non-essential services, such as "1604 air conditioner still not good" as a duplicate task for "1605", the system determines that the status has not been updated, and does not assign another task, only pushes the current status "has been assigned to handle";

[0184] Step 3, self-adaptive adjustment of service cycle, the system sets a new cycle response window for each task:

[0185] Service task Original preset period Period after dynamic adjustment Reason explanation 1602 Replace towel Within 10 minutes Within 15 minutes Robot will queue the delivery task 1605 Air conditioner not cooling Within 10 minutes Keep within 10 minutes Fault type task is given priority 1607 Send water Within 10 minutes Combined with 1603 for 20 minutes Combined task delivery, save round trip time 1603 Send toothbrush Within 10 minutes Combined with 1607 for 20 minutes Same as above 1609 Wall plug power not working (dispatch E02) Within 10 minutes Within 15 minutes Support personnel need to be dispatched across floors 1604 Air conditioner still not working (prompt) Within 10 minutes Pause dispatch Consistent with 1605 status, no need to repeat dispatch 1608 Carpet stains (cleaning) Within 10 minutes Postpone to within 30 minutes Cleaning type task can be queued for non-peak processing

[0186] Customer prompt synchronization (front desk + customer device), the system synchronizes "Your service request has been received and is being processed, please wait patiently", for customers who adjust the processing cycle (such as 1603), the system adds an explanation: "The hotel service is currently busy, and the toothbrush and bottled water delivery tasks are expected to be completed within 20 minutes";

[0187] In summary, in the above example, the system realizes automatic relief of regional load through resource density, service intensity, and periodic adjustment, ensures that emergency tasks are completed first, non-urgent tasks are delayed intelligently, improves service resource utilization efficiency, and reduces repeated services and customer complaints.

[0188] In this embodiment, before the step S1 of classifying the service requests uploaded by the hotel customers to the hotel terminal, the method further comprises:

[0189] S101: identifying the identity label preset by the hotel customer;

[0190] S102: determining whether the identity label belongs to a preset check-in identity;

[0191] S103: if not, limiting the service level of the hotel customer on the hotel terminal based on the identity label, and classifying the available service content of the hotel customer according to the service level.

[0192] In this embodiment, the system identifies the identity label preset by the hotel customer, and then determines whether the identity label belongs to a preset check-in label to perform corresponding steps; for example, when the system determines that the identity label preset by the hotel customer belongs to the preset check-in label, the system considers that the customer has completed the check-in registration, confirms that the customer is a valid check-in customer, and calls the check-in information, preference settings, historical service records, etc. of the customer to ensure service personalization and continuity, grants the customer corresponding room access permissions, facility use permissions, and authorization of related value-added services, and preferentially allocates room services, maintenance responses, and customized needs for the check-in customer to improve customer experience; for example, when the system determines that the identity label preset by the hotel customer does not belong to the preset check-in label, the system considers that the customer has not completed the check-in registration and cannot be confirmed as a valid check-in customer, and limits the service level of the hotel customer on the hotel terminal based on the corresponding identity label and classifies the available service content of the hotel customer according to different service levels; by determining whether the identity label is a valid check-in label, the system can prevent non-registered customers from triggering or misusing service resources, ensure that complete services are only available to check-in customers, reduce service errors or resource waste caused by unknown identities, limit the service level based on the identity label, so that the system can intelligently allocate different customer service resources, response priorities, and service content according to the customer status (such as a reservation customer, a visitor, and a temporary experience user), thereby improving the overall resource scheduling efficiency, and when it is found that the user identity label is not identified as “check-in”, the system can guide the user to complete the check-in registration or registration process in a timely manner, only display basic services (such as consultation and reservation information push), realize service pre-filtering, and improve user conversion rate and platform experience.

[0193] Reference is made to the accompanying drawings Figure 2 In an embodiment of the present application, a hotel front-end and back-end collaborative customer service information flow management system includes:

[0194] A classification module 10 is configured to classify service requests uploaded by hotel customers to a hotel terminal based on information access channels preset by the hotel terminal.

[0195] A judgment module 20 is configured to determine whether the service requests belong to the same service content.

[0196] An execution module 30 is configured to, if yes, identify source information of the service requests, convert the source information into corresponding service mirror information, push the service mirror information to a customer service team preset by the hotel terminal, and generate an autonomous unit of the service requests in real time to the hotel terminal according to a service branch preselected by the customer service team for the service mirror information, wherein the autonomous unit specifically includes an event ID, a location attribute, a request type, and a processing state.

[0197] A second judgment module 40 is configured to determine whether the autonomous unit can match an execution unit preset by the hotel terminal, wherein the execution unit specifically includes a manpower type and a device type.

[0198] A second execution module 50 is configured to, if no, activate an information recognition engine preset by the hotel terminal, construct a fuzzy request corresponding to the autonomous unit through the information recognition engine, divide a service flow of the execution unit according to the fuzzy request, generate a task congestion heat map of the service requests based on a collaborative density and a time lag of the service flow, and dynamically schedule execution resources of the hotel terminal according to the task congestion heat map, wherein the service flow specifically includes a request intention structure, a service task disassembly chain, and an execution path planning, and the task congestion heat map specifically includes a state atlas and a collaborative heat map.

[0199] In the present embodiment, the classification module 10 accesses the information channel pre-set by the hotel terminal based on the information of the hotel terminal, classifies the service requests uploaded by the hotel customers to the hotel terminal, and then the judgment module 20 judges whether the service requests belong to the same service content to perform corresponding steps; for example, when the system determines that the service requests uploaded by the hotel customers to the hotel terminal do not belong to the same service content, the system considers that the multiple service requests currently received have obvious differences in intention, execution mode, resource demand or processing logic, different customers or different rooms submit requests with similar content but different targets, the system constructs an independent service mirror information structure for each different service content request, keeps the logic isolated between requests, avoids confusion, each mirror binds its source, timestamp, context data, at the same time, different service requests are distributed to professional customer service groups to avoid resource conflicts or repeated responses caused by "one order multiple investment", and if multiple different requests come from the same customer (or room) and have no conflict, they can be guided to be combined into "parallel service flow" to improve efficiency, otherwise, they are scheduled in turn according to priority; for example, when the system determines that the service requests uploaded by the hotel customers to the hotel terminal belong to the same service content, the execution module 30 considers that the multiple service requests currently received may have repeated service events, the system identifies the source information of the service requests, converts the source information into corresponding service mirror information, pushes the service mirror information to the pre-selected service branches of the service mirror information in the customer service team pre-set by the hotel terminal, and generates autonomous units of service requests to the hotel terminal in real time, the autonomous unit specifically includes event ID, location attribute, request type and processing state; the system identifies the source information of multiple requests, combines multiple requests belonging to the same service content to generate an autonomous unit, avoids multiple repeated work orders entering the scheduling queue, this deduplication mechanism can prevent human resources and equipment resources from being repeatedly called, especially during peak hours or continuous customer operations (such as repeatedly pressing the service button), which can significantly reduce system congestion rate and execution interference, improve the coordination and execution efficiency of the overall service, at the same time, the original request is converted into standardized service mirror information, which can unify the expression methods of different sources (such as App, voice, front desk), avoid the understanding deviation of customer service or system on service intention, the service mirror contains unified field structure, such as request type, location attribute, timestamp, etc., which lays the information foundation for subsequent generation of autonomous units, ensures that the information received at the execution level is clear, complete and operable, and different customer service teams can pre-set the processing branches (such as delivery, maintenance, cleaning, etc.) of the corresponding service mirror, the system automatically matches the optimal service path according to the mirror content, realizes intelligent classification and accurate distribution of service requests, and the generated autonomous unit not only has a clear event ID and state, but also can be quickly routed to the corresponding post or equipment execution node, thereby realizing highly automated and fine management of service process;Then the second judgment module 40 judges whether these autonomous units can match the execution units pre-set in the hotel terminal. The execution units specifically include human type and equipment type to execute the corresponding steps; for example, when the system determines that the autonomous unit of the service request can match the execution unit pre-set in the hotel terminal, the system will consider that the current task can be identified by the system and find the corresponding execution entity. The system will assign the task to the mobile terminal of the corresponding post personnel (such as the cleaner's PDA, engineering maintenance tablet), including detailed task content, location and response time limit, trigger equipment instructions (such as intelligent delivery robot scheduling, elevator permission opening, air conditioning system automatic response, etc.), and execute specific operations. Update the task status from "pending" to "in execution" or "issued", bind the identifier of the execution unit (such as personnel ID, equipment number), ensure that subsequent tracking is clear and controllable, and continuously monitor the execution feedback of the execution unit, regularly collect execution progress, completion status, abnormal alarms, etc., and feedback to the hotel's central control system or customer service terminal. If necessary, it can support automatic urging or transfer abnormalities to manual processing; for example, when the system determines that the autonomous unit of the service request cannot match the execution unit pre-set by the hotel terminal, the second execution module 50 will think that although the current task can be identified, the corresponding execution entity cannot be found at present. The system will activate the information recognition engine pre-set by the hotel terminal. Through the information recognition engine, the fuzzy request corresponding to the autonomous unit is constructed. Based on these fuzzy requests, the service flow of the execution unit is divided. The service flow specifically includes the request intention structure, the service task disassembly chain and the execution path planning. Based on the collaborative density and time lag of these service flows, the task congestion heat map of the service request is generated. The task congestion heat map specifically includes the state map and the collaborative heat map; the system identifies its potential intention through the information recognition engine and constructs fuzzy requests to avoid the interruption problem of "no execution path after request identification". Through intention understanding and task disassembly, it can further intelligently generate execution plans to fill the gap in the traditional system's ability to handle non-standard tasks. At the same time, by building "service The system can break down ambiguous or complex requests into standard executable subtasks and plan scheduling paths based on priority and accessibility. This approach allows complex tasks to be broken down into multiple concurrent processing units, improving execution efficiency and reducing single-point bottlenecks. Furthermore, combining the collaborative density and time lag of service flows, it generates task congestion heat maps (including state maps and collaborative heat maps) that visually display task distribution and resource pressure. This helps the system implement flow control, detour scheduling, or task rescheduling in high-load areas, preventing resource overload in some areas while leaving others idle. Ultimately, this optimizes resource allocation and balances service response.

[0200] In this embodiment, the execution module further includes:

[0201] A constructing sub-module is configured to construct a clone information object of the service request based on a content structure of the source information, wherein the content structure specifically includes customer identity information, a request original text, a triggering time and a request source, and the clone information object specifically includes a customer communication clone, an execution clone, a tracking clone, a management clone and a multi-language channel adaptation clone;

[0202] A judging sub-module is configured to judge whether states of the clone information objects are uniform.

[0203] An executing sub-module is configured to activate a management mechanism of the clone information object if the states are uniform, to allocate a processing period of the clone information object according to the management mechanism, and to dynamically generate a view aggregator of the clone information object, wherein the management mechanism specifically includes a read-only field, a difference update, an authority control and a version tracking, and the processing period specifically includes a generating, an allocated, a processing, a to-be-confirmed and a closed.

[0204] In the embodiment, the system constructs a clone information object of the service request based on a content structure of the source information, the content structure specifically includes customer identity information, a request original text, a triggering time and a request source, the clone information object specifically includes a customer communication clone, an execution clone, a tracking clone, a management clone and a multilingual channel adaptation clone, and then the system judges whether the states of the clone information objects are uniform to perform corresponding steps; for example, when the system determines that the states of the clone information objects are not uniform, the system considers that a state difference or data asynchronization is generated in multi-link circulation of the current service request, that is, the same customer request presents inconsistent service states or interpretation results in different “clones” such as customer communication, execution layer feedback, state tracking, background management and multilingual translation adaptation, the system compares task IDs, time stamps and execution states in the customer communication clone, the execution clone and the tracking clone to calculate a consistency score; if the score is lower than a threshold value, the system automatically enters a “state reconstruction” process, reinfers an initial service intention from the “request original text + customer identity + triggering time” in the source information, and preferentially allocates a manual customer service if the triggered abnormality is high in severity (such as a safety type or a complaint type), or automatically pulls a standard response path to reissue an execution task if it is a regular service item; for example, when the system determines that the states of the clone information objects are uniform, the system considers that information between each link of the current service request from receiving to execution, tracking, feedback, management and multilingual adaptation is synchronous and consistent, and state drift, interpretation conflict or execution deviation does not occur, the system activates a management mechanism of the clone information object, the management mechanism specifically includes a read-only field, a difference update, a permission control and a version tracking, allocates a processing period of the clone information object according to the management mechanism, and the processing period specifically includes a generated state, an already assigned state, a processing state, a to-be-confirmed state and a closed state.The system can effectively prevent the service request from being tampered or mismodified in the multiple processing processes by activating the management mechanism of the clone information object and introducing the read-only field, differential update, permission control and version tracking. Even if the service request goes through multiple sub-departments and multiple operation nodes, the system can still maintain a highly consistent state view and record the operation history according to the version, thereby improving the transparency and credibility of the system. Meanwhile, the service request is divided into standardized processing cycle stages such as "in generation → assigned → in processing → to be confirmed → closed", which can make the system have clear identification and process control ability for the service state. This not only facilitates the automatic scheduling of human or equipment resources, but also unifies the understanding of the service state by different departments, effectively avoiding misunderstanding, repeated processing or state misplacement. Moreover, each clone object is subject to unified management mechanism, ensuring that its behavior has boundary constraints (such as read-only restriction), dynamic response (such as differential update), access security (such as permission control) and historical traceability (such as version tracking). This helps to control the access and modification of data by different permission levels, accurately identifies abnormal updates or tampering behavior, and supports synchronization and coordination of service requests across languages, roles and platforms.

[0205] In the embodiment, further comprising:

[0206] The generating module is configured to generate a multi-path collaborative distribution instruction of the customer service team based on a task cooperation type of the autonomous unit, where the task cooperation type specifically includes engineering maintenance, cleaning service and room service.

[0207] The third determining module is configured to determine whether the multi-path collaborative distribution instruction can be distributed to idle personnel of the customer service team.

[0208] The third executing module is configured to, if not, list the multi-path collaborative distribution instruction to a preset delay processing queue, dynamically adjust a service path of the multi-path collaborative distribution instruction according to an entry time stamp of the delay processing queue, and adaptively adjust a task time window of the service request according to the service path, where the service path specifically includes a single task, a single task followed by subsequent supplemental distribution and a multi-path collaborative task.

[0209] In the embodiment, the system generates multi-path collaborative distribution instructions of the customer service team based on the task cooperation types of the autonomous units, the task cooperation types specifically including engineering maintenance, cleaning service and room service, and then the system determines whether the multi-path collaborative distribution instructions can be distributed to idle personnel of the customer service team to perform corresponding steps; for example, when the system determines that the multi-path collaborative distribution instructions of the customer service team can be distributed to the idle personnel of the customer service team, the system considers that the current human resource pool is schedulable and available, and the indexes such as skill matching degree, time window and work load between the task and the personnel meet the scheduling conditions, the system considers that the task can be received and executed immediately and efficiently, the system pushes each multi-path collaborative distribution instruction to the terminal of the matched idle personnel, marks the task as “distributed” state, generates a task tracking node in the system for subsequent state collection, feedback and time limit monitoring, at the same time, if the service request corresponds to multiple tasks (for example, a “leakage event” may involve maintenance + cleaning), the system will issue multiple distribution instructions to different team members in parallel, bind them as a whole in the background, and perform overall state normalization management, and can synchronously display “task has been assigned, and the service will be provided within a few minutes” to the customer to improve customer experience and transparency during the waiting process; for example, when the system determines that the multi-path collaborative distribution instructions of the customer service team cannot be distributed to the idle personnel of the customer service team, the system considers that the current human resource pool is not available, and the task and the personnel cannot be matched, the system lists the multi-path collaborative distribution instructions in the pre-set delay processing queue, dynamically adjusts the service path of the multi-path collaborative distribution instructions according to the time stamp of the multi-path collaborative distribution instructions entering the delay processing queue, the service path specifically includes single-path task, single-path task and subsequent supplementary distribution, and multi-path collaborative task, and the task time window of the service request is adaptively adjusted according to the service path; the system avoids blind repeated attempts and scheduling conflicts of the task when the resources are insufficient by adding the task that cannot be immediately assigned to the delay processing queue, and can process the task in order according to the emergency degree and waiting time length by combining the time stamp and the queue rule, so as to ensure that the task execution is more planned and buffered, improve the stability and fairness of the overall resource scheduling, dynamically adjust the original multi-path collaborative task to a single-path task, a supplementary distribution task or a split execution path according to the executable conditions, so that the system can still maintain uninterrupted service when the personnel are insufficient, the adaptive strategy of the task path enhances the flexibility of the task model, effectively improves the response ability of the system in complex scenarios such as peak period, sudden offline of personnel and dense service requests, and the system adaptively updates the task time window of the service request (such as postponing the estimated arrival time, supplementary state notification, etc.) according to the task type and the service path adjustment, avoids the experience loss caused by long-time no feedback of the customer, provides more reasonable buffer time for subsequent execution, maintains the right to know and expectation management of the customer on the task state by cooperating with the customer front-end state synchronization, and improves the overall satisfaction.

[0210] In the embodiment, the second execution module further comprises:

[0211] An acquisition sub-module is configured to acquire virtual intent data of the fuzzy request after the fuzzy request is input into the information recognition engine based on a confidence mechanism preset by the information recognition engine.

[0212] A second judgment sub-module is configured to judge whether the virtual intent data reaches a confidence threshold preset by the confidence mechanism.

[0213] A second execution sub-module is configured to convert the fuzzy request into a preset cold storage request if the virtual intent data does not reach the confidence threshold, retain an intent clue of the cold storage request within a preset time period, collect information keywords repeatedly input into the hotel terminal by the hotel customer, and dynamically wake up the cold storage request according to the same word frequency of the information keywords and the intent clue.

[0214] In the embodiment, the system obtains the virtual intent data of the fuzzy request input into the information recognition engine based on the confidence mechanism preset by the information recognition engine, and then determines whether the virtual intent data reaches the confidence threshold preset by the confidence mechanism to execute corresponding steps; for example, when the system determines that the virtual intent data of the fuzzy request input into the information recognition engine can reach the confidence threshold preset by the confidence mechanism, the system considers that the current fuzzy request can be accurately analyzed by the system to obtain the real intent of the customer, and the system obtains the core fields such as the user intent keyword, target location, and expected time limit from the analysis result, marks the virtual intent as "identified virtual intent", and attaches a confidence value. Meanwhile, the system constructs a corresponding service autonomous unit according to the identified intent, inputs the unit, and still marks it as "generated by fuzzy converted intent", but can be included in the formal scheduling system, and automatically mounts the virtual intent to one or more existing service branches. If there are multiple candidate service branches, the system automatically selects the one with the highest matching degree according to the confidence value and service semantic priority; for example, when the system determines that the virtual intent data of the fuzzy request input into the information recognition engine cannot reach the confidence threshold preset by the confidence mechanism, the system considers that the current fuzzy request cannot be analyzed by the system to obtain the real intent of the customer, and the system converts the fuzzy request into a cold storage request. The intent clues of the cold storage request are preserved for a preset period of time. The information keywords input into the hotel terminal by the hotel customer are collected, and the cold storage request is dynamically awakened according to the same word frequency of the information keywords and the intent clues. When the system cannot clearly identify the intent of the customer, the system does not blindly generate a task, but stores the request in cold storage to prevent resource waste, service deviation, or customer dissatisfaction caused by incorrect scheduling. Through the "intent cold storage + re-determination" mechanism, the system ensures that the service execution is based on high confidence, and ensures that each operation is more targeted and accurate. Meanwhile, the system saves the semantic clues and keyword residual data of the unidentified request, gradually constructs a language preference model of the behavior of the hotel customer, and when the same customer inputs similar expressions in a short period of time, the system can awaken the request based on the "keyword co-occurrence degree" and "context correlation degree" to make the originally unidentified request have context regeneration ability, continuously optimize the understanding of natural language by the system, and establish a fault-tolerant waiting mechanism through the "cold storage request + dynamic awakening" method. Even if the first input fails, the system will not discard it immediately, but wait for the subsequent behavior of the customer for "reconfirmation". This strategy not only improves the tolerance of the system to fuzzy and ambiguous semantics, but also provides a more relaxed operation tolerance for the customer, enhancing the user experience.

[0215] In the embodiment, the determination module further includes:

[0216] The second acquisition submodule is configured to acquire the request interval of the service request based on the space attribute preset by the hotel terminal, where the space attribute specifically includes a room position and a floor position.

[0217] a third judging sub-module, configured to judge whether the request interval is within a preset time window;

[0218] a third executing sub-module, configured to, if yes, identify a trigger frequency of the service request, acquire a group similarity behavior corresponding to the spatial attribute according to the trigger frequency, and put a preset prompt content in a preset public area of the hotel terminal according to the group similarity behavior, where the prompt content is specifically operation suggestions and service mode recommendations.

[0219] In the embodiment, the system obtains the request intervals of the service requests based on the space attributes of the hotel terminal, and the space attributes specifically include room position and floor position. Then, the system determines whether the request intervals are within the pre-set time window to execute corresponding steps. For example, when the system determines that the request intervals of the service requests are not within the pre-set time window, the system considers that the current service requests present non-short-time concentration characteristics in the time dimension, and there is no strong correlation or close trigger logic between the requests. The system considers that the requests have spatial correlation but lack time coincidence, and do not constitute the processing conditions of “task merging” or “batch triggering”. Therefore, the requests should be processed independently according to the respective event IDs, and autonomous units are generated respectively to avoid merging unrelated events into the same service flow, prevent misaggregation, and update the service state nodes of the corresponding rooms or floors in the hotel terminal space attribute graph. The system marks that there are “scattered requests” distributed in the region. If these nodes continue to heat up in a short period of time, the system can set a threshold to trigger regional inspection or preventive task dispatching (such as notifying the cleaning staff and equipment self-checking robots). If the same space region repeatedly appears in the space-time and the content of the service requests is similar (such as multiple customers reporting that the TV cannot be used at different times), the system considers that there is a risk of repeated complaints or an abnormal trend of equipment, and can enter the “abnormal service request aggregation model” to prompt the management personnel to intervene and check. For example, when the system determines that the request intervals of the service requests are within the pre-set time window, the system considers that the requests have obvious concentration or continuity in the time dimension, belong to repeated submission of the same customer, or are prompt requests before the task is completed. The system identifies the trigger frequency of the service requests, obtains the group similarity behavior corresponding to the space attributes according to different trigger frequencies, and puts pre-set prompt contents in the pre-set public areas of the hotel terminal according to the group similarity behavior. The prompt contents specifically include operation suggestions and service mode recommendations.The system automatically judges as possible repeated submission or task urging behavior by identifying the concentration and high frequency of the request, thereby avoiding multiple order placing and repeated processing for the same service event, optimizing resource scheduling efficiency, reducing human waste and device calling redundancy, and improving the intelligent level of the overall service process. At the same time, by analyzing the group similarity behavior, personalized and perceptible prompt information (such as "air conditioner restart needs to wait for 3 minutes, please do not repeat submission", "if you need to speed up the processing, please confirm the service status through the guest room tablet") is put in the public area, which can realize service rhythm guidance without interfering with user privacy, relieve customer anxiety caused by waiting or misoperation, improve service transparency and user experience, and through group behavior analysis (for example, multiple adjacent rooms submit "room temperature abnormality" in the same time period), the system can determine potential device or systematic problems, put in public prompt content in time, trigger background diagnosis tasks, realize proactive intervention and prevention and control before the problem reaches the peak, and enhance the robustness and self-healing ability of the hotel service system.

[0220] In the embodiment, the second determining module further includes:

[0221] A second constructing submodule is configured to construct a processing progress of the service request based on a backhaul processing state of the autonomous unit, where the backhaul processing state specifically includes order accepted, in transit, order completed, and customer confirmation.

[0222] A fourth determining submodule is configured to determine whether the processing progress is stagnant.

[0223] A fourth executing submodule is configured to, if so, acquire a regional resource density of the autonomous unit, dynamically adjust a service intensity of each region of the hotel according to the regional resource density, and adaptively adjust a service cycle of the service request according to the service intensity.

[0224] In the embodiment, the system constructs the processing progress of the service request based on the backhaul processing state of the autonomous unit, and the backhaul processing state specifically includes an order received, in transit, completed, and customer confirmation. Then, the system determines whether the processing progress is stagnant to perform corresponding steps. For example, when the system determines that the processing progress of the service request is not stagnant, the system considers that the current service request is in a normal flow state from dispatch to execution process, and each link of the execution unit is feeding back the processing state according to the process, without abnormal delay or interruption. The system updates the processing record of the service request in real time according to the state of the execution unit (human or device) returned at regular intervals, ensures the traceability of the entire service link, synchronizes the current processing progress to the customer device (such as a room tablet, WeChat applet, etc.) in real time, and prompts the customer to the current state through the form of a "service progress bar" or a "state reminder card", thereby enhancing the perceptibility and satisfaction of the customer during waiting. Moreover, since the processing progress is in a coherent and smooth state, the system does not need to enable an abnormal processing mechanism (such as task reminders, heat map congestion analysis, or replacement mechanism), maintains the automatic execution mode of the current service chain, effectively reduces the management intervention and scheduling resource load, and so on. For example, when the system determines that the processing progress of the service request is stagnant, the system considers that the current service request is in an abnormal flow state. The system acquires the regional resource density of the autonomous unit, dynamically adjusts the service intensity of each region of the hotel according to different regional resource densities, and adaptively adjusts the service period of the service request according to the service intensity. The system can quickly identify the "service bottleneck region" or "human resource shortage area" by analyzing the resource density of the region where the autonomous unit is located (such as the number of execution units in the same region, the availability of human resources, the response frequency of equipment, etc.), avoid blindly global scheduling, reduce the waste of scheduling resources, realize the precise intervention of the abnormal flow request, and so on. According to the service intensity (i.e., the ratio of the aggregation degree of service requests to the task processing capacity in a unit of time), the system can set a more reasonable service time window or appropriately delay the response period for new service requests in the region, so that the processing link of the high-density region obtains a breathing opportunity, guarantees the priority of the existing task closed loop, avoids the influence of long-term suspended tasks on the overall service quality, and so on. The adaptive adjustment of the service period not only improves the sustainability of service execution in the congested area, but also reduces the customer complaint rate caused by task accumulation, enhances the response flexibility of the system to sudden concentrated service requests, and so on. The system controls the processing rhythm of the task in a fine-grained manner, so that the overall hotel service network remains in a "flexible and controllable" state.

[0225] In the embodiment, the system further comprises:

[0226] The recognition module is configured to recognize the identity label preset by the hotel customer.

[0227] The fourth determination module is configured to determine whether the identity label belongs to a preset check-in identity.

[0228] a fourth executing module, configured to: if no, limit a service level of the hotel customer at the hotel terminal based on the identity label, and classify available service contents of the hotel customer according to the service level.

[0229] In the embodiment, the system identifies the identity label pre-set by the hotel customer, and then determines whether the identity label belongs to the pre-set check-in label to perform corresponding steps; for example, when the system determines that the identity label pre-set by the hotel customer belongs to the pre-set check-in label, the system considers that the customer has completed the check-in registration, confirms that the customer is a valid check-in customer, and calls the check-in information, preference settings, historical service records and the like of the customer to ensure service personalization and continuity, and according to the identity label, grants the customer corresponding room access permission, facility use permission and authorization of related value-added services, and preferentially allocates room service, maintenance response and customized needs for the check-in customer to improve customer experience; for example, when the system determines that the identity label pre-set by the hotel customer does not belong to the pre-set check-in label, the system considers that the customer has not completed the check-in registration and cannot confirm that the customer is a valid check-in customer, and limits the service level of the hotel customer at the hotel terminal based on the corresponding identity label, and classifies available service contents of the hotel customer according to different service levels; the system can prevent non-registered customers from triggering or misusing service resources by determining whether the identity label is a valid check-in label, ensure that complete services are only opened to check-in customers, reduce service errors or resource waste caused by unknown identities, limit the service level based on the identity label, so that the system can intelligently allocate different customer service resources, response priorities and service contents according to customer states (such as a reservation customer, a visitor and a temporary experience user), thereby improving overall resource scheduling efficiency, and when it is found that the identity label of a user is not identified as “check-in”, the system can guide the user to complete the check-in registration or registration process in time, only display basic services (such as consultation and reservation information push), realize service pre-filtering, and improve user conversion rate and platform experience.

[0230] Although the embodiments of the present application have been shown and described, it is to be understood that various changes, modifications, substitutions and alterations can be made to the embodiments without departing from the principles and spirit of the present application, and the scope of the present application is defined by the appended claims and their equivalents.

Claims

1. A customer service information flow management method for hotel front-end and back-end collaboration, characterized in that: The following steps are involved: Classifying service requests uploaded by hotel customers to the hotel terminal based on information access channels preset by the hotel terminal; Determining whether the service requests belong to the same service content; If so, identifying the source information of the service request, converting the source information into corresponding service image information, pushing the service image information to a customer service team preset by the hotel terminal, and generating an autonomous unit of the service request to the hotel terminal in real time based on the service branch pre-selected by the customer service team for the service image information, wherein the autonomous unit specifically includes an event ID, a location attribute, a request type, and a processing status; Determining whether the autonomous unit can match an execution unit preset in the hotel terminal, wherein the execution unit specifically includes a human type and an equipment type; If not, the information recognition engine preset by the hotel terminal is activated, and the fuzzy request corresponding to the autonomous unit is constructed through the information recognition engine. According to the fuzzy request, the service flow of the execution unit is divided, and based on the collaborative density and time lag of the service flow, the task congestion heat map of the service request is generated. According to the task congestion heat map, the execution resources of the hotel terminal are dynamically scheduled, wherein the service flow specifically includes a request intention structure, a service task disassembly chain and an execution path planning, and the task congestion heat map specifically includes a state map and a collaborative heat map.

2. The hotel front-end and back-end collaborative customer service information flow management method according to claim 1 is characterized in that: The steps of identifying the source information of the service request, converting the source information into corresponding service image information, and pushing the service image information to a customer service team preset in the hotel terminal further include: Based on the content structure of the source information, a clone information object of the service request is constructed, wherein the content structure specifically includes the customer identity information, the original request text, the trigger time, and the request source, and the clone information object specifically includes a customer communication clone, an execution clone, a tracking clone, a management clone, and a multi-language channel adaptation clone; Determining whether the states of the clone information objects are consistent; If so, activate the management mechanism of the clone information object, allocate the processing cycle of the clone information object according to the management mechanism, and dynamically generate a view aggregator of the clone information object, wherein the management mechanism specifically includes read-only fields, differential updates, permission control and version tracking, and the processing cycle specifically includes generating, assigned, processing, pending confirmation and closed.

3. The hotel front-end and back-end collaborative customer service information flow management method according to claim 1 is characterized in that: After the step of generating the autonomous unit of the service request to the hotel terminal in real time based on the service branch pre-selected by the customer service team based on the service mirror information, the method further includes: generating a multi-channel collaborative distribution instruction for the customer service team based on the task collaboration type of the autonomous unit, wherein the task collaboration type specifically includes engineering maintenance, cleaning service, and room service; Determine whether the multi-channel collaborative distribution instruction can be assigned to an idle person in the customer service team; If not, the multi-way collaborative distribution instruction will be included in the preset delay processing queue, and the service path of the multi-way collaborative distribution instruction will be dynamically adjusted according to the enqueue timestamp of the delay processing queue. Based on the service path, the task time window of the service request will be adaptively adjusted, wherein the service path specifically includes single-way tasks, single-way tasks and subsequent dispatch, and multi-way collaborative tasks.

4. The hotel front-end and back-end collaborative customer service information flow management method according to claim 1 is characterized in that: The step of activating the information recognition engine preset in the hotel terminal and constructing the fuzzy request corresponding to the autonomous unit through the information recognition engine further includes: Based on a confidence mechanism preset by the information recognition engine, obtaining virtual intention data after the fuzzy request is input into the information recognition engine; Determining whether the virtual intention data reaches a confidence threshold preset by the confidence mechanism; If not, the fuzzy request is converted into a preset refrigeration request, the intention clue of the refrigeration request is retained within a preset period of time, the information keywords repeatedly entered by the hotel customer into the hotel terminal are collected, and the refrigeration request is dynamically awakened based on the same word frequency of the information keywords and the intention clues.

5. The hotel front-end and back-end collaborative customer service information flow management method according to claim 1 is characterized in that: The step of determining whether the service requests belong to the same service content further includes: Based on the spatial attributes preset by the hotel terminal, obtaining the request interval of the service request, wherein the spatial attributes specifically include room location and floor location; Determining whether the request interval is within a preset time window; If so, identify the trigger frequency of the service request, obtain the group similarity behavior corresponding to the spatial attribute based on the trigger frequency, and based on the group similarity behavior, place preset prompt content in the public area preset in the hotel terminal, wherein the prompt content specifically includes operation suggestions and service method recommendations.

6. The hotel front-end and back-end collaborative customer service information flow management method according to claim 1 is characterized in that: The step of determining whether the autonomous unit can match the execution unit preset by the hotel terminal further includes: Building a processing progress of the service request based on the feedback processing status of the autonomous unit, wherein the feedback processing status specifically includes received order, in transit, completed, and customer confirmed; Determining whether the processing progress is stagnant; If so, the regional resource density of the autonomous unit is obtained, and the service density of each area of ​​the hotel is dynamically adjusted according to the regional resource density. The service cycle of the service request is adaptively adjusted according to the service density.

7. The hotel front-end and back-end collaborative customer service information flow management method according to claim 1 is characterized in that: Before the step of classifying the service requests uploaded by hotel customers to the hotel terminal based on the information access channels preset by the hotel terminal, the method further includes: Identify the identity tag preset by the hotel customer; Determining whether the identity tag belongs to a preset check-in identity; If not, based on the identity tag, the service level of the hotel customer at the hotel terminal is limited, and the available service content of the hotel customer is classified according to the service level.

8. A hotel front-end and back-end collaborative customer service information flow management system, characterized in that: include: A classification module, configured to classify service requests uploaded by hotel customers to the hotel terminal based on information access channels preset by the hotel terminal; A judgment module, used to judge whether the service requests belong to the same service content; an execution module, configured to, if yes, identify the source information of the service request, convert the source information into corresponding service image information, push the service image information to a customer service team preset by the hotel terminal, and generate an autonomous unit of the service request to the hotel terminal in real time based on the service branch pre-selected by the customer service team for the service image information, wherein the autonomous unit specifically includes an event ID, a location attribute, a request type, and a processing status; A second judgment module is used to judge whether the autonomous unit can match the execution unit preset by the hotel terminal, wherein the execution unit specifically includes a human type and an equipment type; The second execution module is used to activate the information recognition engine preset in the hotel terminal if it cannot, and construct a fuzzy request corresponding to the autonomous unit through the information recognition engine, divide the service flow of the execution unit according to the fuzzy request, and generate a task congestion heat map of the service request based on the collaborative density and time lag of the service flow. According to the task congestion heat map, the execution resources of the hotel terminal are dynamically scheduled, wherein the service flow specifically includes a request intention structure, a service task disassembly chain and an execution path planning, and the task congestion heat map specifically includes a state map and a collaborative heat map.

9. The hotel front-end and back-end collaborative customer service information flow management system according to claim 8 is characterized in that: The execution module also includes: a construction submodule, configured to construct a clone information object of the service request based on the content structure of the source information, wherein the content structure specifically includes customer identity information, original request text, trigger time, and request source, and the clone information object specifically includes a customer communication clone, an execution clone, a tracking clone, a management clone, and a multi-language channel adaptation clone; A judgment submodule, configured to judge whether the states of the clone information objects are consistent; An execution submodule is configured to, if so, activate the management mechanism of the clone information object, allocate the processing cycle of the clone information object according to the management mechanism, and dynamically generate a view aggregator of the clone information object, wherein the management mechanism specifically includes read-only fields, differential updates, permission control, and version tracking, and the processing cycle specifically includes generating, assigned, processing, pending confirmation, and closed.

10. The hotel front-end and back-end collaborative customer service information flow management system according to claim 8, characterized in that: Also includes: a generation module, configured to generate a multi-channel collaborative distribution instruction for the customer service team based on the task collaboration type of the autonomous unit, wherein the task collaboration type specifically includes engineering maintenance, cleaning service, and room service; A third judgment module is used to judge whether the multi-channel coordinated distribution instruction can be assigned to an idle person in the customer service team; The third execution module is used to, if not, list the multi-way collaborative distribution instruction into a preset delay processing queue, dynamically adjust the service path of the multi-way collaborative distribution instruction according to the enqueue timestamp of the delay processing queue, and adaptively adjust the task time window of the service request based on the service path, wherein the service path specifically includes single-way tasks, single-way tasks and subsequent dispatch, and multi-way collaborative tasks.

Images