Order after-sales processing method and device, electronic equipment, storage medium and program product

Abstract

The embodiment of the invention provides an order after-sales processing method and device, electronic equipment, a storage medium and a program product. The method comprises the following steps: in response to obtained order consultation information, calling a return logistics intelligent module to determine a current logistics stage of a to-be-consulted order under the condition that the to-be-consulted order corresponding to the order consultation information is an after-sales scene, any execution strategy in the execution strategy set represents a solution for the order to be consulted in the current logistics stage, and then the execution strategy set matched with the time sequence node corresponding to the current logistics stage is determined according to the time sequence structure of each execution strategy in a pre-maintained execution strategy library, the execution strategies are matched in stages according to the logistics stages, the range of the execution strategies is narrowed, and the accuracy of subsequent solutions is improved; on the basis, analyzing intention information related to the order consultation information; and outputting a solution corresponding to the execution strategy matched with the intention information in the execution strategy set.

Description

Technical Field

[0001] This application relates to the field of computer technology, and in particular to an order after-sales processing method, apparatus, electronic device, storage medium, and program product. Background Technology

[0002] Online customer service refers to the technology that provides answers or consultations to users' inquiries via the internet. Return logistics refers to the logistical stage of a user's order being returned during the after-sales process. In related technologies, for inquiries about return logistics, online customer service typically provides feedback to users based on pre-set response rules with a high degree of relevance.

[0003] In return scenarios, there are usually detailed solutions for specific issues, such as confirming the pickup status and determining the shipping cost. However, pre-set response rules cover a relatively broad range of questions. Generating feedback based solely on these pre-set rules results in poor matching of the questions and is not applicable to the specific circumstances of return logistics, leading to a poor user experience. Summary of the Invention

[0004] To overcome the problems existing in related technologies, embodiments of this application provide an order after-sales processing method, apparatus, electronic device, storage medium, and program product.

[0005] According to a first aspect of the embodiments of this application, an order after-sales processing method is provided, the method comprising: In response to receiving order inquiry information, if the order to be inquired for the order inquiry information is an after-sales scenario, the return logistics intelligent module is invoked to determine the current logistics stage of the order to be inquired for, and the logistics stage is one of at least two preset stages; Based on the temporal structure of each execution strategy in the pre-maintained execution strategy library, a set of execution strategies matching the temporal node corresponding to the current logistics stage is determined; wherein, any execution strategy in the execution strategy set represents a solution for the order to be consulted under the current logistics stage; Analyze the intent information related to the order inquiry; Output the solutions corresponding to the execution strategies that match the intent information in the execution strategy set.

[0006] According to a second aspect of the embodiments of this application, an order after-sales processing apparatus is provided, the apparatus comprising: The logistics stage determination module is used to respond to the acquisition of order inquiry information. If the order to be inquired for the order corresponding to the order inquiry information is an after-sales scenario, the module calls the return logistics intelligent module to determine the current logistics stage of the order to be inquired for the order. The logistics stage is one of at least two preset stages. The execution strategy determination module is used to determine the set of execution strategies that match the time sequence nodes corresponding to the current logistics stage based on the time sequence structure of each execution strategy in the pre-maintained execution strategy library; wherein, any execution strategy in the execution strategy set represents a solution for the order to be consulted under the current logistics stage; The intent determination module is used to parse intent information related to the order inquiry information; The output module is used to output the solutions corresponding to the execution strategies that match the intent information in the execution strategy set.

[0007] According to a third aspect of the present application, an electronic device is provided, including a memory, a processor, and a computer program stored in the memory, the computer program being executed by the processor to cause the electronic device to perform the method as described in the first aspect.

[0008] According to a fourth aspect of the embodiments of this application, a computer-readable storage medium is provided having a computer program stored thereon, the program being executed by a processor to implement the method as described in the first aspect.

[0009] According to a fifth aspect of the embodiments of this application, a computer program product is provided, including instructions that, when executed on a computer, cause the computer to perform the method as described in the first aspect.

[0010] The technical solutions provided in this application embodiment may include the following beneficial effects: The order after-sales processing method in this embodiment, in response to obtaining order consultation information, if the order to be consulted is an after-sales scenario, calls the return logistics intelligent module to determine the current logistics stage of the order to be consulted. The logistics stage is one of at least two preset states. Any execution strategy in the execution strategy set represents a solution for the order to be consulted under the current logistics stage. Then, based on the temporal structure of each execution strategy in the pre-maintained execution strategy library, the set of execution strategies matching the temporal nodes corresponding to the current logistics stage is determined. This, to a certain extent, achieves stage-by-stage matching of execution strategies according to the logistics stage, narrowing the scope of execution strategies and improving the accuracy of subsequent solutions. On this basis, the intent information related to the order consultation information is parsed; and the solution corresponding to the execution strategy in the execution strategy set that matches the intent information is output.

[0011] It should be understood that the above general description and the following detailed description are exemplary and explanatory only, and do not limit this application. Attached Figure Description

[0012] To more clearly illustrate the technical solutions in the embodiments of this application, the accompanying drawings used in the embodiments of this application will be briefly described below. It should be understood that those skilled in the art can obtain other drawings based on these drawings without any creative effort.

[0013] Figure 1 This is a schematic diagram of a scenario corresponding to the order after-sales processing method provided in the embodiments of this application; Figure 2 An exemplary flowchart corresponding to the order after-sales processing method provided in the embodiments of this application; Figure 3 An exemplary method diagram illustrating the order after-sales processing method provided in this application embodiment; Figure 4 An exemplary schematic diagram of the order after-sales processing device provided in the embodiments of this application; Figure 5 This is an exemplary schematic diagram of an order after-sales processing device provided in an embodiment of this application. Detailed Implementation

[0014] The technical solutions of the embodiments of this application will now be described with reference to the accompanying drawings.

[0015] The terminology used in the following embodiments of this application is for the purpose of describing particular embodiments and is not intended to limit the technical solutions of this application. As used in the specification and appended claims of this application, the singular expressions “a,” “an,” “the,” “the,” and “this” are intended to include the plural expressions as well, unless the context clearly indicates otherwise.

[0016] It should also be understood that although the terms "first," "second," etc., may be used in the following embodiments to describe a certain type of object, the objects should not be limited to these terms. These terms are used to distinguish the specific implementation objects of that type of object.

[0017] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, stored data, displayed data, etc.) involved in the embodiments of this application are all information and data authorized by the user or fully authorized by all parties. Furthermore, the collection, use and processing of related data must comply with the relevant laws, regulations and standards of the relevant countries and regions, and corresponding operation entry points are provided for users to choose to authorize or refuse.

[0018] The technical scenarios of the embodiments of this application are described below.

[0019] Online customer service refers to a service that leverages internet technology to provide real-time answers or remote support for various user inquiries and questions. In e-commerce, after-sales service, and other fields, online customer service has become a crucial channel for user communication. Return logistics, as a key link in the after-sales process, involves the entire logistics process from when a user submits a return request to when the goods are shipped back to the merchant or warehouse. In related technical solutions, for user-raised return logistics questions, online customer service systems typically rely on a pre-set response rule library. Through semantic matching or keyword recognition, they select rules with high matching degrees to generate standard responses and provide feedback to the user.

[0020] However, in specific return scenarios, the questions users ask are often highly specific and contextualized. For example, users may need to clarify specific operational questions such as "When will the package be picked up after being sent?", "Does the store support door-to-door pickup?", "Who bears the return shipping costs?", and "How do I fill in the tracking number?". These questions not only involve the logistics stage but also multiple dimensions such as operational guidelines, responsibilities, and cost attribution, requiring a comprehensive judgment based on specific information such as order status, logistics information, and platform policies.

[0021] Currently, rule-based response mechanisms, due to their broad coverage of questions, offer only general responses that lack adaptation to specific return scenarios. For example, an online customer service system might recognize the keyword "return logistics" but fail to differentiate whether the user is inquiring about "progress," "fees," or "operation methods," resulting in a discrepancy between the generated response and the user's actual needs. This lack of matching not only reduces problem-solving efficiency but also degrades the user experience, failing to meet their expectations for accurate and efficient after-sales service.

[0022] In view of this, embodiments of this application propose an order after-sales processing method, which can be applied to an order after-sales processing system. For example... Figure 1 As shown, this is an order after-sales processing system that implements the order after-sales processing method. Figure 1 The illustrated order after-sales processing system includes: an order confirmation status unit 101, an after-sales unit 102, and a return unit 103. Each unit in this system can be a software module, an intelligent agent, a hardware circuit, or a combination of software and hardware with specific functions, used to execute the methods described below to achieve precise and automated processing of after-sales scenarios such as return logistics.

[0023] The order status confirmation unit 101 serves as the data foundation and entry point for the order after-sales processing system. It is configured to respond to user after-sales requests and determine the current status of the target order. Its specific tasks include: accessing the user's inquiry session and retrieving the corresponding target order based on the user's identifier or order identifier; then, obtaining real-time status information of the order from the order center, logistics platform, etc., such as whether the order has been shipped, whether it has been dispatched, the current node of the logistics trajectory, and whether it has been signed for. This unit ensures that subsequent processing flows can be based on accurate and real-time order data, providing crucial input for precise scenario judgment.

[0024] The after-sales unit 102 is configured to determine whether to allow the more efficient return unit 103 to process the order based on its current status. The after-sales unit 102 can be pre-configured with processing logic covering various after-sales scenarios, and is used to determine the relevant order inquiries that require return logistics in after-sales statuses such as refund only or return with refund, and then transfer them to the return unit 103 for processing.

[0025] The return unit 103 can be the execution terminal of the order after-sales processing system, communicatively connected to the after-sales unit 102. This return unit 103 is configured to determine and execute a target execution strategy based on the logistics stage of the return order, generating and providing after-sales guidance information (solution) to the user that is appropriate to the current status of the target order. After determining the target execution strategy, the return unit 103 will dynamically execute a series of operations included in that logic. For example, for the "return logistics - pending shipment" scenario, its target processing logic might include: automatically displaying the nearest self-pickup point address to the user, generating a pickup appointment link, clearly stating who bears the shipping costs, and providing an entry point for the tracking number.

[0026] The return order 103 may further include a pre-processing subunit 1031, a strategy set determination subunit 1032, a target execution strategy determination subunit 1033, and an execution subunit 1034. The pre-processing subunit 1031 further confirms whether the order inquiry information corresponds to the return logistics information, providing multiple verifications for after-sales processing. If not, it is transferred to other after-sales related procedures. If so, it is further handed over to the strategy set determination subunit 1032 to determine the execution strategy set based on the logistics stage. The target execution strategy subunit 1033 determines the corresponding target execution strategy from the execution strategy set based on the user session context and order-related information. Finally, the execution subunit 1034 executes the target execution strategy.

[0027] By working together with the order status confirmation unit 101, after-sales unit 102, and return unit 103, the efficiency and completion rate of automated processing are improved, overcoming the problems of answer generalization and low matching degree caused by matching based on a fixed rule base in related technologies.

[0028] It should be understood that Figure 1 The schematic diagram of the application scenario is only an illustrative representation of the order after-sales processing system involved in the embodiments of this application, and does not constitute a limitation on the technical solutions of the embodiments of this application.

[0029] In the above operating environment, the embodiments of this application provide as follows: Figure 2 The order after-sales processing method is shown below. Please refer to it. Figure 2 This order after-sales processing method is applied to, for example... Figure 1 The return unit 103 in the order details. The after-sales processing method for this order includes the following steps: In step S201, in response to obtaining order consultation information, if the order to be consulted corresponding to the order consultation information is an after-sales scenario, the return logistics intelligent module is invoked to determine the current logistics stage of the order to be consulted.

[0030] The logistics stage is one of at least two pre-set stages.

[0031] For example, order consultation information can be a complete context data packet when a user initiates a consultation request, including user identification, order identification, consultation text, product snapshot, current page tracking information, historical dialogue records, and other data. The after-sales scenario refers to a task domain strongly related to the user's consultation intent and the after-sales service chain of a completed order, covering service requests after order completion such as returns and refunds, exchanges and repairs, shipping disputes, and logistics anomalies. The intelligent return logistics module can be integrated into the agent in the dialogue system. This agent encapsulates a logistics status recognition engine, a shipping logistics rule tree decision engine, and multi-standard operating procedure (SOP) collaborative scheduling capabilities, specifically handling consultation issues in the return logistics chain. Logistics status refers to the discrete stage labels of return logistics on the fulfillment timeline, which can include atomic states such as pre-pickup and post-pickup.

[0032] The system analyzes received order inquiries to determine whether the orders in question fall under after-sales service. This determination is achieved by identifying keywords in the inquiry information, such as detecting the presence of words related to after-sales service like "return," "exchange," and "repair." The algorithm employs a Keyword Matching Algorithm (KMA), which matches the inquiry text against a pre-defined database of after-sales keywords. If a match is found, the order is classified as an after-sales service. This approach accurately distinguishes the type of order inquiry, filtering out orders in after-sales service scenarios for subsequent processing by a dedicated intelligent return logistics module, improving efficiency and targeting.

[0033] After determining that the order to be consulted falls under an after-sales scenario, the intelligent return logistics module is invoked. This module acts as a dedicated "assistant" for handling return logistics issues, participating in the current order consultation process. Based on the relevant information of the order to be consulted, the intelligent return logistics module interacts with the logistics system to query the current logistics stage of the order. The logistics stage may include various situations such as pre-collection and post-collection. For example, by connecting with the courier company's logistics data interface, it obtains real-time information such as the order's collection time and transportation progress, thereby determining whether the order is in the pre-collection stage awaiting courier pickup or has already been collected and is en route.

[0034] In step S202, based on the temporal structure of each execution strategy in the pre-maintained execution strategy library, a set of execution strategies that matches the temporal node corresponding to the current logistics stage is determined.

[0035] In this set of execution strategies, any execution strategy represents a solution for the order to be consulted at the current logistics stage.

[0036] For example, the timing structure of each execution strategy is pre-maintained in the intelligent return logistics module. This timing structure can be a timeline or decision tree where each execution strategy in the execution strategy library is organized according to the actual task flow of the return logistics, creating a timeline with dependencies between strategies. This structure ensures the correspondence between strategies and logistics stages. For example, the timing structure could be: User Application → Merchant Approval → User Ships → Courier Pickup → In Transit → Merchant Signs.

[0037] The execution strategy can be a document recording solutions to all possible problems at the corresponding logistics stage. These solutions can be in text or table format. For example, a text document could be created for the pre-pickup stage of the logistics process. This document could include all possible questions users might ask before pickup. For instance, regarding urging the courier to pick up the package, it could be divided into two scenarios: whether the scheduled pickup has expired. If the scheduled pickup has expired, the execution strategy could be set as: changing the courier company or urging the courier to pick up the package.

[0038] Once the current logistics stage is determined, it can be determined based on the correspondence between the identifiers of each execution strategy set and the logistics stage.

[0039] In step S203, the intent information related to the order inquiry information is parsed.

[0040] For example, intent information can refer to the core service requests or action instructions abstracted and summarized from the user's order inquiry information, revealing the essence of what the user wants or hopes to do.

[0041] Upon receiving an order inquiry, it's not enough to analyze the inquiry text in isolation. A more comprehensive understanding can be achieved by combining the logistics stage and the user's historical behavior data. For example, if the order inquiry is "What should I do now?", the intent might be unclear on its own. However, by considering the known "logistics stage" ("awaiting shipment") and the order status ("return request approved"), the user's intent can be understood within the context of "awaiting shipment."

[0042] For example, pre-trained intent recognition models or large language models can be used for intent recognition. Before recognition, a pre-defined after-sales intent tag library can be established. Intent information is determined based on indicators such as the confidence level of each tag in the after-sales intent tag library, along with order inquiry information. The intent tag library can include queries about door-to-door pickup appointment methods, urging pickup, inquiries about freight payment methods, confirmation of correct tracking number entries, and queries about real-time logistics location.

[0043] In step S204, the solution corresponding to the execution strategy that matches the intent information is output from the execution strategy set.

[0044] For example, after determining the execution strategy, the corresponding solution is sent to the client. The solution may include a solution title, a solution preview, solution details, and the solution execution result.

[0045] As can be seen, in response to receiving order consultation information, when the order consultation information corresponds to an after-sales scenario, the intelligent return logistics module is invoked to determine the current logistics stage of the order. The logistics stage is one of at least two preset states. Each execution strategy in the execution strategy set represents a solution for the order under the current logistics stage. Furthermore, the temporal structure of each execution strategy in the pre-maintained execution strategy library is used to determine the execution strategy set that matches the temporal node corresponding to the current logistics stage. This, to a certain extent, achieves stage-by-stage matching of execution strategies based on the logistics stage, narrowing the scope of execution strategies and improving the accuracy of subsequent solutions. Based on this, the intent information related to the order consultation information is parsed; and the solution corresponding to the execution strategy in the execution strategy set that matches the intent information is output.

[0046] In some embodiments, determining the set of execution strategies can also be achieved by: determining the timing node corresponding to the current logistics stage; determining at least one execution strategy whose timing is later than the timing node and belongs to the end node of the current logistics stage based on the timing structure of each execution strategy; and determining the set of execution strategies from the at least one execution strategy.

[0047] For example, in the "time sequence structure" of return logistics, a time sequence node represents a specific, discrete task state point. It is the logical representation of the logistics stage in the strategy library and has a clear sequential relationship with other nodes in the time sequence structure. For example, "the courier has completed pickup" can be a time sequence node. An end node refers to the time sequence node corresponding to the end point of the current logistics stage. It is not the end point of the entire return process, but rather the boundary of the logistics stage currently under discussion. For example, for the "in transit" stage, its "end node" might be "the package has arrived at the destination distribution center" or "in delivery," marking the completion of the "trunk transportation" process.

[0048] After determining the logistics stage, an inverted index based on logistics stage identifiers can be built for the execution strategy library, mapping each logistics stage identifier to an associated time-series node. For example: before pickup → time-series node A, after pickup → time-series node B. This can accelerate the retrieval speed from status identifier to time-series node. Input the current logistics stage identifier into the inverted index to retrieve all matching time-series node identifiers.

[0049] Next, the sequence structure is traversed forward (downstream) to determine the boundary of the current logistics stage. The execution strategies corresponding to all nodes from the current sequence node to the end node are compiled into an execution strategy set. This boundary is the end node. The rules for identifying the end node are predefined; for example, it can be set that encountering the first node belonging to the next logistics stage, or encountering a specific status flag, indicates that it is the "end node" of the previous stage. This is essentially a directed graph traversal or rule matching process.

[0050] If multiple execution strategies correspond to a single time-series node, the strategies in the execution strategy set can be sorted according to preset rules. If conflicting strategies exist, the superior strategy is selected through weighted scoring or a task rule engine. These preset rules can be set according to criteria such as cost, timeliness, and user preferences.

[0051] For example, strategy A has low cost but high risk, while strategy B has high cost but high accuracy. If the user is a member, strategy B should be preferred; if the user is a regular user, strategy A should be preferred. This ensures that the strategy set conforms to relevant logic while also meeting differentiated needs. For example, high-value orders automatically trigger manual review to reduce the risk of loss.

[0052] like Figure 3 The diagram shown is an example of an execution strategy library. Figure 3 The root node represents the return shipment corresponding to the return logistics, and different execution strategies apply to different shipment methods. Figure 3Taking the return logistics corresponding to door-to-door pickup as an example, it can be further divided into pre-pickup, post-pickup, and no-pickup status based on whether the shipping order for door-to-door pickup has been created and whether door-to-door pickup has been carried out. Assuming the current logistics stage corresponds to pre-pickup, then... Figure 3 In the tree structure, locate the time sequence node corresponding to "Returned Goods Logistics - Before Pickup", and use the execution strategies after that time sequence node as the execution strategy set (the corresponding execution strategies). Figure 3 (End node not shown).

[0053] It should be noted that, Figure 3 The grayscale groups shown represent the execution strategies implemented by the intelligent agent, dynamically deciding on pre-collection requests such as "door-to-door pickup" reminders and contract rescheduling, and supporting flexible persuasive guidance for "order cancellation". The non-grayscale groups correspond to traditional SOP execution strategies.

[0054] After determining the set of execution strategies corresponding to the current logistics stage, in order to improve the response speed to order inquiry information, when determining the target execution strategy, the execution strategies in the execution strategy set can be run separately based on the execution rules matching the order to be inquired about, and the running results of each execution strategy can be obtained; any running result indicates whether the corresponding execution strategy is available; from the multiple execution strategies with available running results, the execution strategies that meet the matching degree of the order inquiry information are determined as the execution strategy set.

[0055] For example, the execution rule can be a set of admission conditions in each SOP strategy, which can be composed of Boolean expressions to verify the feasibility of the strategy in the current order context, covering dimensions such as order status constraints, user qualification verification, merchant capability matching, and external system availability checks.

[0056] The execution strategy is loaded sequentially according to its original priority. The rule parser is then invoked to convert the execution rule text in the strategy document into an abstract syntax tree, extracting the variable names and operators that the rules depend on. Simultaneously, a data source mapping table is constructed to determine whether each variable should be obtained from the order center, user profiling service, or courier company. This step transforms the task rules described in natural language into machine-executable structured expressions, improving processing speed.

[0057] Based on the aforementioned data source mapping table, an asynchronous data acquisition pipeline is started. The corresponding data sources are obtained by batch calling tools such as the order query interface, user tag service, and logistics details. Of course, the acquired data sources can also be processed by means of completion, normalization, etc., to unify the data format.

[0058] The processed data source is input into the corresponding target execution strategies to obtain the execution results for each strategy. The execution results can be whether the target execution strategy is executable and information after executing the strategy; alternatively, the execution results can indicate that the data source cannot meet the conditions of one or more strategies within the target execution strategy. For example, the target execution strategies before pickup are: Change courier company → Company A, Company B, and Company C. Due to insufficient capacity or other reasons, Company C cannot be selected when changing courier companies. Therefore, only the strategies "Change courier company → Company A" and "Company B" are executable, while "Change courier company → Company C" is not.

[0059] If the execution results of multiple target execution strategies are all usable, then a matching degree calculation can be performed based on the strategy representation vector of each usable strategy and the feature vector of the order consultation information. This matching degree calculation is only used for the calculation of usable target execution strategies, reducing the waste of computing resources. After obtaining the matching degree score of each usable target execution strategy, they are sorted in descending or ascending order of score, and the usable target execution strategy with the highest matching degree score is selected as the target execution strategy.

[0060] In some embodiments, determining the intent information related to order consultation information can also be achieved by: extracting semantic features representing the order consultation information; identifying whether the semantic features match preset intent tags; if the semantic features match preset intent tags, determining the intent tags that match the semantic features as intent information; if the semantic features do not match preset intent tags, invoking the dialogue intelligence module to engage in dialogue with the user, and extracting intent information based on the collected dialogue information in response to the collected dialogue information; wherein, the dialogue information includes inquiry information or execution strategies.

[0061] For example, a semantic coding model can be used to process the text of order consultation information, specifically, converting the text into semantic features. These semantic features are then classified; they can be input into an intent classifier to obtain the probability distribution of intent labels corresponding to each semantic feature. The intent information is then determined based on this probability distribution. During the determination of intent information, a confidence score can be generated based on this probability distribution. If the confidence score is greater than a preset confidence threshold, a clear intent information is obtained; if the confidence score is less than the preset confidence threshold, it indicates that no intent label has been matched, and further inquiry information needs to be sent to the user to confirm the intent information.

[0062] For example, the dialogue intelligence module is invoked to engage in dialogue with the user. In response to the collected dialogue information, intent information is extracted based on the dialogue information; the dialogue information includes inquiry information or execution strategies.

[0063] When the dialogue information is execution strategy information, the dialogue intelligence module is invoked to display multiple execution strategies to the user; in response to receiving a selection operation for any of the multiple execution strategies, the intent information is determined according to the execution strategy corresponding to the selection operation, and the solution corresponding to the execution strategy corresponding to the selection operation is output.

[0064] For example, the dialogue information sends confirmation messages to the user regarding various execution strategies. For instance, changing the courier company and urging door-to-door pickup are both available target execution strategies before pickup, requiring confirmation of the user's request. For example, the inquiry message could be, "Dear customer, we apologize that the courier was late. Would you like us to urge them on time, or would you like us to switch to a better courier service?" If the user's response based on the inquiry message clearly indicates "urge them on time," then "urging door-to-door pickup" is taken as the intent information, and a solution for the corresponding execution strategy is output. If the user's response based on the inquiry message clearly indicates "change the courier service," then "change the courier company" is taken as the intent information, and a solution for the corresponding execution strategy is output.

[0065] When the dialogue information is an inquiry, an inquiry message is sent to the dialogue client based on a preset inquiry script template; wherein, the dialogue client is a dialogue client that provides order consultation information; in response to the reply information of the inquiry message received from the dialogue client, the intent information to be confirmed is extracted; in response to the confirmation instruction of the intent to be confirmed is received, the intent information to be confirmed is determined as intent information.

[0066] For example, the preset inquiry script templates can be a predefined, structured text generation rule base. Each inquiry script template is associated with one or more uncertain intent scenarios, and its content is designed to guide users to clarify their true needs. The inquiry script templates typically contain placeholders that can be dynamically populated based on the current context (such as logistics stage, list of suspected intents).

[0067] First, based on the context of the current conversation and the order's logistics stage, the most suitable template is selected from a pre-set query template library. Then, specific intent options are dynamically populated into the template's placeholders to generate a query message. These intent options can be among the top intent tags with high confidence, such as checking logistics progress or inquiring about shipping rules. Finally, this message is pushed to the current conversation client and displayed to the user.

[0068] Receive the response information returned from the dialogue client. Analyze the response information. Since the query information is highly structured, the user's response information is usually also highly structured, possibly imperative or a direct restatement of options. Using keyword matching or pattern matching algorithms, map the response content back to one of the alternative intents provided in the query information, and mark this intent as intent information to be confirmed. If the user's response is unexpectedly complex, semantic analysis can also be initiated to confirm it.

[0069] Upon receiving the intent information to be confirmed, it is not immediately treated as the final intent. To ensure no misunderstanding during the interaction, a confirmation step is executed. For example, a confirmation message containing a summary of the intent information to be confirmed is sent to the user, and a confirmation reply is awaited. A positive reply constitutes a confirmation instruction. For instance, asking, "Okay, you want to check the logistics progress, right?" and receiving replies like "Yes" or "Yes." If the confidence level of the response is extremely high, the intent information to be confirmed can be directly elevated to the final confirmed intent information. This "secondary confirmation" mechanism improves the accuracy of intent recognition and the user experience.

[0070] In some embodiments, there are intent information corresponding to multiple execution strategies, that is, a set of execution strategies. Therefore, after determining the intent information, it is also possible to: determine the matching degree between each execution strategy in the set of execution strategies and the intent information; and push multiple execution strategies to the user in order of matching degree from high to low.

[0071] For example, if the number of execution strategies corresponding to the intent information is small, the corresponding execution strategy can be directly pushed to the user. During the push process, the execution strategies can be pushed sequentially according to their matching degree with the intent information, with the execution strategies with a higher matching degree being pushed to the user first.

[0072] If the execution strategy set corresponding to the intent information contains a large number of execution strategies, directly pushing all execution strategies in the set to the user would be wasteful of resources and degrade the user experience. Therefore, a subset of execution strategies can be selected and pushed to the chat client based on their matching degree with the intended intent. Similarly, they can be pushed sequentially according to their matching degree.

[0073] For example, if a user only expresses "the package hasn't been picked up yet," which actually describes a problem, then all available target execution strategies can be output to the user. The output order can be determined according to the matching degree calculated above, so that the user can determine the corresponding target execution strategy according to their needs later. For example, determine the matching degree of multiple target execution strategies with the order consultation information respectively; and push multiple target execution strategies to the user in order of matching degree from high to low.

[0074] In some embodiments, outputting the solution corresponding to the target execution strategy includes: for any execution strategy in the execution strategy set, parsing the execution path contained in the target execution strategy and the description of the execution actions contained in the execution path; generating a solution based on the execution path and the description of the execution actions contained in the execution path.

[0075] For example, an execution path is a structured processing flow defined by the target execution strategy itself, also known as a Standard Operating Procedure (SOP). The execution path clarifies the sequential stages required to resolve a user's problem. For instance, the execution path for a "urging door-to-door collection" strategy might be as follows: soothe the user's emotions → verify key information → perform the urging action → promise feedback to the user → provide follow-up guidance.

[0076] Action descriptions can correspond to one or more specific "action descriptions" at each stage of the execution path. These descriptions specify the specific tasks to be completed or the specific content to be generated or displayed to the user at that stage. For example, in the stage of verifying key information, action descriptions might include: "Extract the 'latest scheduled pick-up time' from the order data" or "Determine if the current time has expired"; in the stage of executing the expedited action, action descriptions might include: "Call the logistics company's expedited order interface and pass in the order number" or "Record the expedited order request serial number and timestamp."

[0077] Generating a solution is not simply about throwing out a strategy document; rather, it involves dynamically and structurally constructing a solution package that can be directly executed by the dialogue system based on the parsed execution path and action description. The content generation process can involve generating specific, personalized text, voice, or other multimedia responses based on the interaction-related parts of the action description. For example, combining the script template "We're sorry the courier didn't arrive on time; we've urged them again" from the action description with specific order information (such as order number and appointment time) generates the final sentence for the user: "Dear customer, we're sorry your order ending in XXXX was scheduled for pickup at XX:00 today, but it's past the appointed time. We've contacted the courier again to urge them on; please wait a moment." Furthermore, relevant links or action entry points can be embedded based on the action description. For example, if the solution includes "providing users with a link to check logistics progress," the tracking link for the current order will be automatically generated and embedded in the response. If the solution needs to guide the user to make a selection, an interactive menu with options will be generated.

[0078] The final generated solution can be a structured object or data package, which can be user-facing response content, background execution instruction set, next step logic judgment node, etc. The next step logic judgment node can indicate the next step of the dialogue flow or task logic after the current solution is executed, such as waiting for user response, jumping to another strategy, or ending the current service.

[0079] The solution is generated based on in-depth analysis and dynamic assembly of the target execution strategy, rather than simply filling in a static template. This ensures that the output solution not only accurately corresponds to the user's current inquiry and order status, but is also an actionable, logical, and personalized complete processing plan that can efficiently guide the dialogue or task flow to complete the problem-solving, improving response speed and the consistency of the service experience.

[0080] There are also some embodiments that generate a solution based on the execution path and the execution action description contained in the execution path, including: after executing the execution action of the I-th node on the execution path, determining whether the I-th node is the last node of the execution path; if the I-th node is the last node, then outputting the result of the I-th node as the solution; if the I-th node is not the last node, then continuing to execute the I+1-th node until the I-th node is the last node.

[0081] For example, the system locates the I-th node on the execution path and executes all the actions defined by that node. For instance, if the action of the I-th node is "call the logistics company's order reminder interface," a request is sent to the specified reminder interface, and a return result is obtained, such as "order reminder successful" or "order reminder failed, reason: courier busy." Key inputs, outputs, and state changes during execution are recorded in real-time within the context of the current session, providing a data foundation for subsequent node judgments and the generation of the final solution. This ensures that the tasks of each node are executed accurately and that their results are retained, allowing the task flow to progress based on actual execution conditions rather than relying on static presets.

[0082] After the current node's action is completed, it is determined whether the current I-th node is the "last node" defined in the entire execution path. This is typically achieved by comparing the unique identifier of the current node (such as a step ID or sequence number) with the endpoint identifier marked in the execution path metadata. This is a critical routing decision point, ensuring the integrity of the process. If the I-th node is the last node, the results and data generated by all nodes in this execution path (from the first to the I-th node) are integrated to generate a solution.

[0083] In the previous embodiments, the logistics stage could be pre-pickup and post-pickup. In addition, if the user cancels the shipping order after confirming the return or the merchant refuses the return, there is no shipping information at this time, which is the post-cancellation status. Based on the above embodiments, before determining the execution strategy set corresponding to the current logistics stage, it also includes: determining whether the processing status of the order to be consulted is the post-cancellation status; if the processing status of the order to be consulted is the post-cancellation status, then the logistics stage is set to no shipping information.

[0084] For example, the detailed status record of the after-sales order corresponding to the "order pending consultation" identifier can be obtained. The latest status change information, such as "closed," "user cancelled," or "merchant rejected," can be used to ensure the accuracy of logistics stage determination.

[0085] If the current order is in the "Cancelled After-Sales" status, any historical or cached logistics information that might be retrieved from the logistics company's interface will be ignored, and the "Logistics Stage" variable will be directly set to "No Shipment Information." Setting the logistics stage to "No Shipment Information" prevents accidental operations. This ensures that solutions are not provided based on old logistics information generated by an outdated after-sales process, thus avoiding confusion and errors for both users and delivery companies. For example, if a user cancels a return, the system might still be urging the courier to pick up a cancelled package.

[0086] As can be seen, this application's embodiments achieve a refined and intelligent upgrade of strategy matching in return logistics consultation scenarios, improving problem-solving efficiency and solution accuracy. Specifically, this solution reduces the scope of strategy matching from global search to local precise matching through pre-identification of the logistics stage and phased isolation of the strategy set, reducing unnecessary resource waste; it accurately determines user intent by combining semantic matching and historical dialogue information. Simultaneously, by pre-running the target execution strategy in the execution strategy set, it achieves real-time prediction of strategy availability and on-demand resource loading, improving response speed. At the user experience level, the dynamic solution generation mechanism based on the execution path can provide personalized dialogue and interactive guidance, thereby improving user satisfaction and reducing the rate of repeated consultations due to ambiguous intent. Furthermore, the tree-structured, time-series strategy library makes strategy management scalable and traceable, reducing the new task access cycle.

[0087] The above implementation steps can also be implemented through software modules to achieve the corresponding functions. Corresponding to the above order after-sales processing method, this application embodiment can also provide an order after-sales processing device.

[0088] like Figure 4As shown, an order after-sales processing device is provided, which is applied to a target device. The device may include: a logistics stage determination module 41, an execution strategy determination module 42, an intent determination module 43, and an output module 44.

[0089] For example: the logistics stage determination module 41 is used to, in response to receiving order consultation information, call the return logistics intelligent module to determine the current logistics stage of the order in question when the order consultation information corresponds to an after-sales scenario, wherein the logistics stage is one of at least two preset stages; the execution strategy determination module 42 is used to determine the set of execution strategies that match the time sequence node corresponding to the current logistics stage based on the temporal structure of each execution strategy in the pre-maintained execution strategy library; wherein any execution strategy in the execution strategy set represents a solution for the order in question under the current logistics stage; the intent determination module 43 is used to parse the intent information related to the order consultation information; and the output module 44 is used to output the solution corresponding to the execution strategy in the execution strategy set that matches the intent information.

[0090] In some optional embodiments, the execution strategy determination module is further configured to: Determine the time sequence node corresponding to the current logistics stage; Based on the temporal structure of each execution strategy, determine at least one execution strategy whose temporal sequence is later than the temporal node and belongs to the end node of the current logistics stage. The set of execution strategies is determined from the at least one execution strategy.

[0091] In some optional embodiments, the execution strategy determination module is further configured to: Based on the execution rules matching the orders to be consulted, the execution strategies are executed in a centralized manner to obtain the execution results of each execution strategy; any execution result indicates whether the corresponding execution strategy is available; The execution strategy set is determined from the multiple execution strategies that are available after the execution result, and the matching degree of the order consultation information meets the condition.

[0092] In some optional embodiments, the intent determination module is further configured to: extract semantic features representing the order inquiry information; Identify whether the semantic features match a preset intent label; If the semantic feature matches a preset intent tag, the intent tag matched by the semantic feature is determined as the intent information; If the semantic features do not match the preset intent tags, the dialogue intelligence module is invoked to engage in dialogue with the user. In response to the collected dialogue information, the intent information is extracted based on the dialogue information; wherein, the dialogue information includes inquiry information or execution strategies.

[0093] In some optional embodiments, when the dialogue information is execution policy information, the intent determination module is further configured to: The dialogue intelligence module is invoked to display multiple execution strategies to the user; In response to receiving a selection operation for any of the multiple execution strategies, the intent information is determined based on the execution strategy corresponding to the selection operation, and the solution corresponding to the execution strategy corresponding to the selection operation is output.

[0094] In some optional embodiments, when the dialogue information is an inquiry, the intent determination module is further configured to: send the inquiry information to the dialogue client based on the preset inquiry script template; wherein, the dialogue client is the dialogue client that provides the order inquiry information; In response to the inquiry information received from the dialogue client, extract the intent information to be confirmed; In response to receiving a confirmation instruction for the intent to be confirmed, the intent information to be confirmed is determined as the intent information.

[0095] In some optional embodiments, the intent determination module, after determining the intent label matched by the semantic features as the intent information, is further configured to: Determine the matching degree between each execution strategy in the execution strategy set and the intent information; Multiple execution strategies are pushed to the user in descending order of matching degree.

[0096] In some optional embodiments, the output module is further configured to: parse the execution path contained in the target execution strategy and the execution action description contained in the execution path; The solution is generated based on the execution path and the description of the execution actions contained in the execution path.

[0097] In some optional embodiments, the output module is further configured to: determine whether the I-th node is the last node of the execution path after executing the execution action of the I-th node on the execution path; If the I-th node is the last node, then output the result of the I-th node as the solution; If the I-th node is not the last node, then continue executing the (I+1)-th node until the I-th node becomes the last node.

[0098] In some optional embodiments, the at least two states include pre-pickup, post-pickup, and no shipment information. Before determining the set of execution strategies corresponding to the current logistics stage, it is also used to: determine whether the processing status corresponding to the order to be consulted is the post-cancellation status. If the processing status of the order to be consulted is the post-sales cancellation status, then the logistics stage is set to no shipment information.

[0099] As can be seen, in the embodiments of this application, in response to obtaining order consultation information, when the order to be consulted is an after-sales scenario, the return logistics intelligent module is invoked to determine the current logistics stage of the order to be consulted. The logistics stage is one of at least two preset states. Any execution strategy in the execution strategy set represents a solution for the order to be consulted under the current logistics stage. Then, the temporal structure of each execution strategy in the pre-maintained execution strategy library is used to determine the execution strategy set that matches the temporal node corresponding to the current logistics stage. To a certain extent, this achieves stage-by-stage matching of execution strategies according to the logistics stage, narrows the scope of execution strategies, and improves the accuracy of subsequent solutions. On this basis, the intent information related to the order consultation information is parsed; and the solution corresponding to the execution strategy in the execution strategy set that matches the intent information is output.

[0100] Understandable, Figure 4 The division of the various modules is merely a logical functional division. In actual implementation, the functions of these modules can be integrated into the hardware entity of the electronic device.

[0101] Please refer to Figure 5 , Figure 5 An electronic device is provided, and this disclosure also provides an electronic device for performing some or all of the steps of the above-described order after-sales processing method. Please refer to... Figure 5 The electronic device includes a processor 500, a memory 501, a bus 502, and a communication interface 503, wherein the processor 500, the communication interface 503, and the memory 501 are connected via the bus 502; the memory 501 stores a computer program that can run on the processor 500, and the processor 500 can be equivalent to a CPU, executing the aforementioned provisions of this disclosure when running the computer program. Figure 2 The order after-sales processing method provided in any of the embodiments, and in the process of executing the aforementioned method, interacts with the corresponding client through the communication interface 503 to exchange data and instructions.

[0102] The memory 501 may include high-speed random access memory (RAM) or non-volatile memory (NVM), such as at least one disk storage device. Communication between this system network element and at least one other network element is achieved through at least one communication interface 503 (which can be wired or wireless), such as the Internet, wide area network, local area network, metropolitan area network, etc.

[0103] Bus 502 can be an Industry Standard Architecture (ISA) bus, a Peripheral Component Interconnect (PCI) bus, or an Extended Industry Standard Architecture (EISA) bus, etc. The bus can be divided into an address bus, a data bus, a control bus, etc. Memory 501 is used to store programs, and the processor 500 executes the programs after receiving execution instructions. Figure 2 The illustrated implementation shows that the order after-sales processing method can be applied to the processor 500, or implemented by the processor 500.

[0104] The processor 500 may be an integrated circuit chip with signal processing capabilities. In implementation, each step of the above method can be completed by the integrated logic circuitry in the hardware of the processor 500 or by instructions in software form. The processor 500 may be a general-purpose processor, including a CPU, a network processor (NP), etc.; it may also be a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field-programmable gate array (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, or discrete hardware components. It can implement or execute the methods, steps, and logic block diagrams disclosed in the embodiments of this disclosure. The general-purpose processor may be a microprocessor or any conventional processor. The steps of the methods disclosed in the embodiments of this disclosure can be directly embodied in the execution of a hardware decoding processor, or executed by a combination of hardware and software modules in the decoding processor. The software modules may reside in random access memory, flash memory, read-only memory, programmable read-only memory, electrically erasable programmable memory, registers, or other mature storage media in the art. The storage medium is located in memory 501. The processor 500 reads the information in memory 501 and, in conjunction with its hardware, completes the steps of the above method.

[0105] The electronic device provided in this disclosure and the order after-sales processing method provided in this disclosure are based on the same inventive concept and have the same beneficial effects as the methods they adopt, operate or implement.

[0106] This application also provides a computer-readable storage medium storing instructions for order after-sales processing, which, when run on a computer, cause the computer to perform some or all of the steps in the method described in the foregoing embodiments.

[0107] This application also provides a computer program product that includes instructions for order after-sales processing, which, when run on a computer, causes the computer to perform some or all of the steps in the method described in the foregoing embodiments.

[0108] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working processes of the systems, devices, and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0109] In the several embodiments provided in this application, it should be understood that the disclosed systems, apparatuses, and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be an indirect coupling or communication connection between apparatuses or units through some interfaces, and may be electrical, mechanical, or other forms.

[0110] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0111] Furthermore, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0112] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, smartphone, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0113] Although alternative embodiments of this application have been described, those skilled in the art, upon learning the basic inventive concept, can make further changes and modifications to these embodiments. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments as well as all changes and modifications falling within the scope of this application.

[0114] The specific embodiments described above further illustrate the purpose, technical solution, and beneficial effects of this application. It should be understood that the above description is only a specific embodiment of this application and is not intended to limit the scope of protection of this application. Any modifications, equivalent substitutions, improvements, etc., made on the basis of the technical solution of this application should be included within the scope of protection of this invention.

Claims

1. A method for processing after-sales orders, characterized in that, The method includes: In response to receiving order inquiry information, if the order to be inquired for the order inquiry information is an after-sales scenario, the return logistics intelligent module is invoked to determine the current logistics stage of the order to be inquired for, and the logistics stage is one of at least two preset stages; Based on the temporal structure of each execution strategy in the pre-maintained execution strategy library, a set of execution strategies matching the temporal node corresponding to the current logistics stage is determined; wherein, any execution strategy in the execution strategy set represents a solution for the order to be consulted under the current logistics stage; Analyze the intent information related to the order inquiry; Output the solutions corresponding to the execution strategies that match the intent information in the execution strategy set.

2. The method of claim 1, wherein, The determination of the set of execution strategies matching the time sequence node corresponding to the current logistics stage includes: Determine the time sequence node corresponding to the current logistics stage; Based on the temporal structure of each execution strategy, determine at least one execution strategy whose temporal sequence is later than the temporal node and belongs to the end node of the current logistics stage. The set of execution strategies is determined from the at least one execution strategy.

3. The method of claim 2, wherein, Determining the set of execution strategies from the at least one execution strategy includes: Based on the execution rules matching the orders to be consulted, the execution strategies are executed in a centralized manner to obtain the execution results of each execution strategy; any execution result indicates whether the corresponding execution strategy is available; The execution strategy set is determined from the multiple execution strategies that are available after the execution result, and the matching degree of the order consultation information meets the condition.

4. The method according to any one of claims 1-3, characterized in that, The process of parsing the intent information related to the order inquiry information includes: Extract the semantic features represented by the order inquiry information; Identify whether the semantic features match a preset intent label; If the semantic feature matches a preset intent tag, the intent tag matched by the semantic feature is determined as the intent information; If the semantic features do not match the preset intent tags, the dialogue intelligence module is invoked to engage in dialogue with the user. In response to the collected dialogue information, the intent information is extracted based on the dialogue information; wherein, the dialogue information includes inquiry information or execution strategies.

5. The method according to claim 4, characterized in that, When the dialogue information is execution strategy information, the step of extracting the intent information based on the dialogue information in response to the acquisition of the dialogue information includes: The dialogue intelligence module is invoked to display multiple execution strategies to the user; In response to receiving a selection operation for any of the multiple execution strategies, the intent information is determined based on the execution strategy corresponding to the selection operation, and the solution corresponding to the execution strategy corresponding to the selection operation is output.

6. The method according to claim 4, characterized in that, When the dialogue information is an inquiry, the response to collecting the dialogue information, extracting the intent information based on the dialogue information, includes: Based on the preset inquiry script template, an inquiry message is sent to the dialogue client; wherein, the dialogue client is the dialogue client that provides the order inquiry information; In response to the inquiry information received from the dialogue client, extract the intent information to be confirmed; In response to receiving a confirmation instruction for the intent to be confirmed, the intent information to be confirmed is determined as the intent information.

7. The method according to claim 4, characterized in that, After determining the intent label matched by the semantic features as the intent information, the method further includes: Determine the matching degree between each execution strategy in the execution strategy set and the intent information; Multiple execution strategies are pushed to the user in descending order of matching degree.

8. The method according to claim 1, characterized in that, The step of outputting the solution corresponding to the execution strategy that matches the intent information in the execution strategy set includes: For any execution strategy in the execution strategy set, parse the execution path contained in the execution strategy and the execution action description contained in the execution path; The solution is generated based on the execution path and the description of the execution actions contained in the execution path.

9. The method according to claim 8, characterized in that, The step of generating the solution based on the execution path and the description of the execution actions contained in the execution path includes: After executing the action of the I-th node on the execution path, determine whether the I-th node is the last node on the execution path; If the I-th node is the last node, then output the result of the I-th node as the solution; If the I-th node is not the last node, then continue executing the (I+1)-th node until the I-th node becomes the last node.

10. The method according to claim 1, characterized in that, The at least two stages include pre-pickup, post-pickup, and no shipment information. Before determining the set of execution strategies corresponding to the current logistics stage, the process also includes: Determine whether the processing status of the order to be consulted is a post-sales cancellation status; If the processing status of the order to be consulted is the post-sales cancellation status, then the logistics stage is set to no shipment information.

11. An order after-sales processing device, characterized in that, The device includes: The logistics stage determination module is used to respond to the acquisition of order inquiry information. If the order to be inquired for the order corresponding to the order inquiry information is an after-sales scenario, the module calls the return logistics intelligent module to determine the current logistics stage of the order to be inquired for the order. The logistics stage is one of at least two preset stages. The execution strategy determination module is used to determine the set of execution strategies that match the time sequence nodes corresponding to the current logistics stage based on the time sequence structure of each execution strategy in the pre-maintained execution strategy library; wherein, any execution strategy in the execution strategy set represents a solution for the order to be consulted under the current logistics stage; The intent determination module is used to parse intent information related to the order inquiry information; The output module is used to output the solutions corresponding to the execution strategies that match the intent information in the execution strategy set.

12. An electronic device comprising a memory, a processor, and a computer program stored in the memory, wherein the processor, when executing the computer program, implements the method of any one of claims 1-10.

13. A computer-readable storage medium having a computer program stored thereon, characterized in that, The program is executed by a processor to implement the method as described in any one of claims 1-10.

14. A computer program product, characterized in that, Includes instructions that, when executed on a computer, cause the computer to perform the method as described in any one of claims 1-10.

Images